]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
26 years agoAdd entries for xl0, tlc0 and adv0. Some of these aren't even
Jordan K. Hubbard [Mon, 7 Sep 1998 13:00:58 +0000 (13:00 +0000)]
Add entries for xl0, tlc0 and adv0.  Some of these aren't even
in LINT!

26 years agoPut the zombie ffs sysctl node in "notyet" state together with its few
Bruce Evans [Mon, 7 Sep 1998 11:50:19 +0000 (11:50 +0000)]
Put the zombie ffs sysctl node in "notyet" state together with its few
remaining children.  Prepare it for MOUNT_UFS going away.

26 years agoUpdate alpha configuration files and remove i386 target from
Doug Rabson [Mon, 7 Sep 1998 08:18:34 +0000 (08:18 +0000)]
Update alpha configuration files and remove i386 target from
libbinutils/config.h - it is declared in the arch Makefile.

26 years agoUse Elf_Addr to store addresses instead of Elf_Word.
Doug Rabson [Mon, 7 Sep 1998 07:30:44 +0000 (07:30 +0000)]
Use Elf_Addr to store addresses instead of Elf_Word.

26 years agoDevice framework code now declared in MI code.
Doug Rabson [Mon, 7 Sep 1998 07:29:30 +0000 (07:29 +0000)]
Device framework code now declared in MI code.

26 years agoFix problem reported on bugtraq: check permission of device mounted
Guido van Rooij [Mon, 7 Sep 1998 07:20:30 +0000 (07:20 +0000)]
Fix problem reported on bugtraq: check permission of device mounted
for non-root users. Fortunately, the default for vfs.usermount is 0.
Tested by: "Jan B. Koum " <jkb@best.com

26 years agoMake MFS do the default on VOP_FREEBLKS().
Poul-Henning Kamp [Mon, 7 Sep 1998 06:52:01 +0000 (06:52 +0000)]
Make MFS do the default on VOP_FREEBLKS().

XXX: we could deallocate the storage, but somebody else will
have to pick up that task.

26 years agoSet the ${suffix} only once...
Andrzej Bialecki [Mon, 7 Sep 1998 06:46:25 +0000 (06:46 +0000)]
Set the ${suffix} only once...

26 years ago* Add code to display statistics continuously, every some seconds.
Andrzej Bialecki [Mon, 7 Sep 1998 06:41:14 +0000 (06:41 +0000)]
* Add code to display statistics continuously, every some seconds.
* Fix a number of mem leaks.
* Update README.

Submitted by: Luigi Rizzo <luigi@iet.unipi.it>

26 years agoInstall the legacy man file in section 1aout.
John Birrell [Mon, 7 Sep 1998 06:01:37 +0000 (06:01 +0000)]
Install the legacy man file in section 1aout.

26 years agoAdd a man section 1aout for the legacy man files to avoid them being
John Birrell [Mon, 7 Sep 1998 05:55:02 +0000 (05:55 +0000)]
Add a man section 1aout for the legacy man files to avoid them being
installed over the ones from binutils.

26 years agoUse a default execution type if none of the listed ones match. This
Sean Eric Fagan [Mon, 7 Sep 1998 05:49:43 +0000 (05:49 +0000)]
Use a default execution type if none of the listed ones match.  This
uses the first type listed in the array as the default type.  This isn't
perfect, but I thought it would be better than nothing.

26 years agoInstall the legacy man file in section 1aout.
John Birrell [Mon, 7 Sep 1998 05:46:09 +0000 (05:46 +0000)]
Install the legacy man file in section 1aout.

26 years agoAdd directories for the legacy tool man files to stop them over-writing
John Birrell [Mon, 7 Sep 1998 05:43:41 +0000 (05:43 +0000)]
Add directories for the legacy tool man files to stop them over-writing
the ones from binutils.

26 years agoMade unloading of the nfs LKM sort of work. This is mainly to test
Bruce Evans [Mon, 7 Sep 1998 05:42:15 +0000 (05:42 +0000)]
Made unloading of the nfs LKM sort of work.  This is mainly to test
detachment of vfs sysctls.  Unloading of vfs LKMs doesn't actually
work for any vfs, since it leaves garbage pointers to memory
allocation control structures.

26 years agoDon't go below the low water mark of free pages due to optional prefaulting
Tor Egge [Sun, 6 Sep 1998 23:04:20 +0000 (23:04 +0000)]
Don't go below the low water mark of free pages due to optional prefaulting
of pages.
PR: 2431

26 years agoUpdate for import of binutils-2.9.1.
John Polstra [Sun, 6 Sep 1998 23:01:53 +0000 (23:01 +0000)]
Update for import of binutils-2.9.1.

Submitted by: Doug Rabson <dfr>

26 years agoResolve conflicts from import of binutils-2.9.1.
John Polstra [Sun, 6 Sep 1998 23:00:35 +0000 (23:00 +0000)]
Resolve conflicts from import of binutils-2.9.1.

Submitted by: Doug Rabson <dfr>

26 years agoThis commit was generated by cvs2svn to compensate for changes in r38889,
John Polstra [Sun, 6 Sep 1998 22:57:45 +0000 (22:57 +0000)]
This commit was generated by cvs2svn to compensate for changes in r38889,
which included commits to RCS files with non-trunk default branches.

26 years agoImport GNU binutils-2.9.1. This will break things for a few minutes
John Polstra [Sun, 6 Sep 1998 22:57:45 +0000 (22:57 +0000)]
Import GNU binutils-2.9.1.  This will break things for a few minutes
until I've made the commits to resolve the conflicts.

Submitted by: Doug Rabson <dfr>

26 years agoMaintain a mapping from irq number to (ioapic number, int pin) tuple,
Tor Egge [Sun, 6 Sep 1998 22:41:42 +0000 (22:41 +0000)]
Maintain a mapping from irq number to (ioapic number, int pin) tuple,
and use this when masking/unmasking interrupts.

Maintain a mapping from (iopaic number, int pin) tuple to irq number,
and use this when configuring devices and programming the ioapics.

Previous code assumed that irq number was equal to int pin number, and
that the ioapic number was 0.

Don't let an AP enter _cpu_switch before all local apics are initialized.

26 years agoBetter handling of word splitting. Don't record the same region
Tor Egge [Sun, 6 Sep 1998 21:13:09 +0000 (21:13 +0000)]
Better handling of word splitting.  Don't record the same region
multiple times when performing nested variable expansion, and
preserve some quoting information in order to avoid removing
apparently empty expansion result.

26 years agoDon't blindly eliminate `..' and the previous pathname component.
Tor Egge [Sun, 6 Sep 1998 21:01:57 +0000 (21:01 +0000)]
Don't blindly eliminate `..' and the previous pathname component.
PR: 2541
Obtained from: NetBSD

26 years agoFix calls to mmap. It returns void *, and on failure it returns
John Polstra [Sun, 6 Sep 1998 20:43:25 +0000 (20:43 +0000)]
Fix calls to mmap.  It returns void *, and on failure it returns
MAP_FAILED.

Don't try to extend the mapping in place if it is too short.
There's no guarantee it will be possible.  Remap the file instead.

Put in a few style fixes.

Submitted by: Bruce Evans <bde>

26 years agoClean LINT
Robert V. Baron [Sun, 6 Sep 1998 20:01:34 +0000 (20:01 +0000)]
Clean LINT

26 years agoPatch the transmit error handler to avoid following NULL pointers and
Bill Paul [Sun, 6 Sep 1998 14:58:46 +0000 (14:58 +0000)]
Patch the transmit error handler to avoid following NULL pointers and
generating a trap 12 panic. The code blindly assumed that in the event
of a transmit error, the packet that caused the error would still be
at the head of the driver's transmit queue (sc->xl_cdata.xl_tx_head).
However in the case of error 82 (which indicates that a transmit error
occurred after part of the transmit FIFO memory has been reclaimed)
this is not true: the TX queue has already been flushed, and the
pointer to the head of the queue is NULL, so trying to dereference
the pointer to find the transmit descriptor address causes a crash.

The code now checks for a NULL pointer before trying to reload the
chip's download pointer register. There may still be error messages
printed warning of the transmit error, but no panic should occur.

Note that this eror code is only generated with "cyclone" chipsets
(3c900B, 3c905B, and presumeably the 3c980 server adapter). It should
only appear during periods of heavy traffic, probably only on
non-switched networks.

Problem reported by: Darcy Buskermolen <darcy@ok-connect.com>

26 years agoAdd a compuserve example.
Brian Somers [Sun, 6 Sep 1998 13:55:51 +0000 (13:55 +0000)]
Add a compuserve example.
Detailed by: Masahiro Sekiguchi <seki@sysrap.cs.fujitsu.co.jp>

26 years agoFix tons of cut-n-paste errors in the WANT_NFS3 code.
David E. O'Brien [Sun, 6 Sep 1998 09:21:35 +0000 (09:21 +0000)]
Fix tons of cut-n-paste errors in the WANT_NFS3 code.

26 years agoReviewed by: Andrey A. Chernov <ache@nagual.pp.ru>, Doug Rabson <dfr@nlsystems.com>
Matthew Dillon [Sun, 6 Sep 1998 09:11:06 +0000 (09:11 +0000)]
Reviewed by: Andrey A. Chernov <ache@nagual.pp.ru>, Doug Rabson <dfr@nlsystems.com>

    Rewrote sleep() to remove int/long typing assumptions between the argument
    to the function and the timespec structure.

26 years agoUse a varargs style prototype in the manual page, with accompanying
Joseph Koshy [Sun, 6 Sep 1998 08:50:49 +0000 (08:50 +0000)]
Use a varargs style prototype in the manual page, with accompanying
explanation in the text.

26 years agotcp_extensions now only applies to RFC1323
Poul-Henning Kamp [Sun, 6 Sep 1998 08:20:11 +0000 (08:20 +0000)]
tcp_extensions now only applies to RFC1323

26 years agoRFC 1644 has the status "Experimental Protocol", which means:
Poul-Henning Kamp [Sun, 6 Sep 1998 08:17:35 +0000 (08:17 +0000)]
RFC 1644 has the status "Experimental Protocol", which means:

   4.1.4.  Experimental Protocol

      A system should not implement an experimental protocol unless it
      is participating in the experiment and has coordinated its use of
      the protocol with the developer of the protocol.

Pointed out by: Steinar Haug <sthaug@nethelp.no>

26 years agoStore formatted panic string in static buffer to make it available later
Andrey A. Chernov [Sun, 6 Sep 1998 06:25:18 +0000 (06:25 +0000)]
Store formatted panic string in static buffer to make it available later
for savecore.
Previous code give only panic format to savecore

26 years agoChange the ELF64_R_SYM() to shift 32 bits instead of 8 bits to match
John Birrell [Sat, 5 Sep 1998 23:07:59 +0000 (23:07 +0000)]
Change the ELF64_R_SYM() to shift 32 bits instead of 8 bits to match
binutils.

Ensure that three of the structures are the size that binutils writes
them. I just love code that doesn't share header files to avoid
problems like these.

With this change rtld-elf works on alpha.

26 years agoNFS can use TCP now days
David E. O'Brien [Sat, 5 Sep 1998 20:42:47 +0000 (20:42 +0000)]
NFS can use TCP now days

26 years agoDon't recognize a file as an a.out shared library unless it has at
John Polstra [Sat, 5 Sep 1998 20:28:48 +0000 (20:28 +0000)]
Don't recognize a file as an a.out shared library unless it has at
least 2 version numbers.  This fixes the bug where the dynamic
linker would try to load an ELF shared library if it found one.

Note, this change also fixes the same thing in "ld", because the
code is shared.

For "ld" there is still a problem with ".a" libraries, which cannot
be distinguished by name.  I haven't decided what, if anything, to
do about that.

26 years agoIgnore the statically configured vfs type numbers and assign vfs
Bruce Evans [Sat, 5 Sep 1998 17:13:28 +0000 (17:13 +0000)]
Ignore the statically configured vfs type numbers and assign vfs
type numbers in vfs attach order (modulo incomplete reuse of old
numbers after vfs LKMs are unloaded).  This requires reinitializing
the sysctl tree (or at least the vfs subtree) for vfs's that support
sysctls (currently only nfs).  sysctl_order() already handled
reinitialization reasonably except it checked for annulled self
references in the wrong place.

Fixed sysctls for vfs LKMs.

26 years agoFor a.out mode, ignore shared libraries that don't have at least 2
John Polstra [Sat, 5 Sep 1998 16:20:15 +0000 (16:20 +0000)]
For a.out mode, ignore shared libraries that don't have at least 2
version numbers.

26 years agoDon't do anything special to prevent access to created files - honour
Bruce Evans [Sat, 5 Sep 1998 15:56:10 +0000 (15:56 +0000)]
Don't do anything special to prevent access to created files - honour
the default umask.  World unreadable files broke installing over nfs.

26 years agoInstantiate `nfs_mount_type' in a standard file so that it is present
Bruce Evans [Sat, 5 Sep 1998 15:17:34 +0000 (15:17 +0000)]
Instantiate `nfs_mount_type' in a standard file so that it is present
when nfs is an LKM.  Declare it in a header file.  Don't forget to use
it in non-Lite2 code.  Initialize it to -1 instead of to 0, since 0
will soon be the mount type number for the first vfs loaded.

NetBSD uses strcmp() to avoid this ugly global.

26 years agoDisable gcc's builtin memcpy for alpha since it doesn't cope with unaligned
Doug Rabson [Sat, 5 Sep 1998 15:11:47 +0000 (15:11 +0000)]
Disable gcc's builtin memcpy for alpha since it doesn't cope with unaligned
regions properly and this triggers an unaligned access trap.

26 years agoFixed bogotification of pseudocode for syscall args by rev.1.53 of
Bruce Evans [Sat, 5 Sep 1998 14:30:11 +0000 (14:30 +0000)]
Fixed bogotification of pseudocode for syscall args by rev.1.53 of
syscalls.master.

26 years agoSysctl nodes are written to, so don't put them in the text section.
Bruce Evans [Sat, 5 Sep 1998 14:13:35 +0000 (14:13 +0000)]
Sysctl nodes are written to, so don't put them in the text section.

Our write protection of the kernel text on i386's doesn't actually
work in many cases:
- use of the 4MB page completely breaks it.
- CR0_WP isn't set until just before init is started, so the kernel
  text is not write protected during kernel initialization.

26 years agoAdd a new vnode op, VOP_FREEBLKS(), which filesystems can use to inform
Poul-Henning Kamp [Sat, 5 Sep 1998 14:13:12 +0000 (14:13 +0000)]
Add a new vnode op, VOP_FREEBLKS(), which filesystems can use to inform
device drivers about sectors no longer in use.

Device-drivers receive the call through d_strategy, if they have
D_CANFREE in d_flags.

This allows flash based devices to erase the sectors and avoid
pointlessly carrying them around in compactions.

Reviewed by: Kirk Mckusick, bde
Sponsored by: M-Systems (www.m-sys.com)

26 years agoFixed recently perpetrated printf format errors.
Bruce Evans [Sat, 5 Sep 1998 13:24:39 +0000 (13:24 +0000)]
Fixed recently perpetrated printf format errors.

26 years agoDon't comment out dead code. Remove it.
Bruce Evans [Sat, 5 Sep 1998 12:42:56 +0000 (12:42 +0000)]
Don't comment out dead code.  Remove it.

Fixed disgustingly long lines.

Improved English in some comments.

26 years agoRemove tcl from src/contrib.
Satoshi Asami [Sat, 5 Sep 1998 12:33:01 +0000 (12:33 +0000)]
Remove tcl from src/contrib.

26 years agoRemove tcl from /usr/src, according to prior agreement. BTW, tcl-8.0 *is*
Satoshi Asami [Sat, 5 Sep 1998 12:25:55 +0000 (12:25 +0000)]
Remove tcl from /usr/src, according to prior agreement.  BTW, tcl-8.0 *is*
a port so there is nothing to be done on that side now.

Approved by: jkh
===
To: "Jordan K. Hubbard" <jkh@time.cdrom.com>
cc: Andreas Klemm <andreas@klemm.gtn.com>, current@freebsd.org
Subject: Re: Make this a relese coordinator decision (was Re: ports-current/packages-current discontinued)
From: David Greenman <dg@root.com>
Date: Sun, 03 Aug 1997 20:23:31 -0700

>decision is, I'll respect it.

   Another chance to architect people's principles...I can hardly wait. Seems
quite appropriate for a Sunday - I just need to get one of those collection
plates (and money envelopes) so I can profit, too. :-)
   Tcl stays in /usr/src for now, but it needs to be kept up to date; same
for perl. If Jordan doesn't have "setup" (written in tcl) ready for 3.0,
then tcl will be yanked prior to the 3.0 release (and made into a port).
   As for the ports tree only supporting the last FreeBSD release, this seems
sensible to me. The "ports" have always been a moving target between releases
and the problem is only going to get worse when we expand to supporting other
processor architectures. In any case, Satoshi is and always has been in charge
of the ports tree and whatever he wants to do with it (within reason :-)) is
his decision.
   Does this cover the issue completely? I admit to deleting messages in this
thread with unusual fervor (people have FAR too much time on their hands!).
There's a fair bit of reasoning behind the above, but since everyone is sick
of arguing about this, I'll spare you the analysis.

-DG

David Greenman
Core-team/Principal Architect, The FreeBSD Project

26 years agoChange i386 in a few paths to ${MACHINE} to support MACHINE=pc98.
John Birrell [Sat, 5 Sep 1998 11:37:14 +0000 (11:37 +0000)]
Change i386 in a few paths to ${MACHINE} to support MACHINE=pc98.

26 years agoChange MACHINE to MACHINE_ARCH when testing for i386 to support
John Birrell [Sat, 5 Sep 1998 08:35:53 +0000 (08:35 +0000)]
Change MACHINE to MACHINE_ARCH when testing for i386 to support
MACHINE=pc98.

There are a lot unsupported machines here.

26 years agoChaneg MACHINE to MACHINE_ARCH to support MACHINE=pc98.
John Birrell [Sat, 5 Sep 1998 08:33:10 +0000 (08:33 +0000)]
Chaneg MACHINE to MACHINE_ARCH to support MACHINE=pc98.

Remove a couple of unsupported machines.

26 years agoChaneg MACHINE to MACHINE_ARCH to support MACHINE=pc98.
John Birrell [Sat, 5 Sep 1998 08:31:10 +0000 (08:31 +0000)]
Chaneg MACHINE to MACHINE_ARCH to support MACHINE=pc98.

26 years agoChange MACHINE to MACHINE_ARCH to support MACHINE=pc98.
John Birrell [Sat, 5 Sep 1998 08:29:05 +0000 (08:29 +0000)]
Change MACHINE to MACHINE_ARCH to support MACHINE=pc98.

26 years agoAdd support for MACHINE=pc98. This leaves a single legacy entry for
John Birrell [Sat, 5 Sep 1998 08:27:56 +0000 (08:27 +0000)]
Add support for MACHINE=pc98. This leaves a single legacy entry for
the old aout bootloader.

26 years agoChange MACHINE to MACHINE_ARCH to support MACHINE=pc98.
John Birrell [Sat, 5 Sep 1998 08:22:30 +0000 (08:22 +0000)]
Change MACHINE to MACHINE_ARCH to support MACHINE=pc98.

26 years agoAdd a GENERICupgrade config file for pc98 users to upgrade to elf
John Birrell [Sat, 5 Sep 1998 08:14:36 +0000 (08:14 +0000)]
Add a GENERICupgrade config file for pc98 users to upgrade to elf
in the same way as i386 users. This is a copy of GENERIC98.

26 years agoChange a hardcoded i386 reference to ${MACHINE} so that MACHINE=pc98
John Birrell [Sat, 5 Sep 1998 08:10:44 +0000 (08:10 +0000)]
Change a hardcoded i386 reference to ${MACHINE} so that MACHINE=pc98
is supported properly.

26 years agoUse INT_MAX instead of LONG_MAX since the variable being compared
John Birrell [Sat, 5 Sep 1998 08:01:26 +0000 (08:01 +0000)]
Use INT_MAX instead of LONG_MAX since the variable being compared
is an int, not a long.

26 years agoChange a test for NETBSD_SYSCALLS to __alpha__. We're not ready to
John Birrell [Sat, 5 Sep 1998 07:57:43 +0000 (07:57 +0000)]
Change a test for NETBSD_SYSCALLS to __alpha__. We're not ready to
do profiling on alpha yet.

26 years agoChange the profile data file name from gmon.out to progname.gmon
John Birrell [Sat, 5 Sep 1998 07:56:36 +0000 (07:56 +0000)]
Change the profile data file name from gmon.out to progname.gmon
where progname comes from __progname in crt0 or crt1.

26 years agoSet CHROOTDIR's initial object type also.
Jordan K. Hubbard [Sat, 5 Sep 1998 07:43:13 +0000 (07:43 +0000)]
Set CHROOTDIR's initial object type also.
Suggested by: jb

26 years agoNow that ldconfig does elf work too, build it on alpha as well.
John Birrell [Sat, 5 Sep 1998 07:35:19 +0000 (07:35 +0000)]
Now that ldconfig does elf work too, build it on alpha as well.

26 years agoMore sorting of references.
David E. O'Brien [Sat, 5 Sep 1998 06:56:36 +0000 (06:56 +0000)]
More sorting of references.

26 years agoAdd OBJFORMAT to list of variables passed into /mk script since it seems
Jordan K. Hubbard [Sat, 5 Sep 1998 06:54:42 +0000 (06:54 +0000)]
Add OBJFORMAT to list of variables passed into /mk script since it seems
to think you're using aout otherwise (and emits a rather confusing and
inaccurate diagnostic about OBJFORMAT being set in /etc/make.conf when
it most demonstrably is not set there).

26 years agoMFO: rev 1.3 - correct man page xrefs.
David E. O'Brien [Sat, 5 Sep 1998 06:38:19 +0000 (06:38 +0000)]
MFO: rev 1.3 - correct man page xrefs.

26 years agocompare getopt() with -1, not EOF
David E. O'Brien [Sat, 5 Sep 1998 06:06:00 +0000 (06:06 +0000)]
compare getopt() with -1, not EOF

26 years agoMFO rev 1.4 - sort cross references
David E. O'Brien [Sat, 5 Sep 1998 05:52:24 +0000 (05:52 +0000)]
MFO rev 1.4 - sort cross references

26 years agoExplicitly pass the "-aout" option to ldconfig, until I've had a
John Polstra [Sat, 5 Sep 1998 04:27:35 +0000 (04:27 +0000)]
Explicitly pass the "-aout" option to ldconfig, until I've had a
chance to work out a better scheme to handle the separate
invocations necessary for a.out and ELF.

26 years agoImplement ldconfig functionality for ELF. The hints are stored in
John Polstra [Sat, 5 Sep 1998 03:31:00 +0000 (03:31 +0000)]
Implement ldconfig functionality for ELF.  The hints are stored in
a different file than the a.out hints, namely, "/var/run/ld-elf.so.hints".
These hints consist only of the directory search path.  There is
no hash table as in the a.out hints, because ELF doesn't have to
search for the file with the highest minor version number.  (It
doesn't have minor version numbers at all.)

A single run of ldconfig updates either the a.out hints or the ELF
hints, but not both.  The set of hints to process is selected in
the usual way, via /etc/objformat, or ${OBJFORMAT}, or the "-aout"
or "-elf" command line option.  The rationale is that you probably
want to search different directories for ELF than for a.out.

"ldconfig -r" is faked up to produce output like we are used to,
except that for ELF there are no minor version numbers.  This should
enable "ldconfig -r" to be used for checking LIB_DEPENDS in ports
even for ELF.

I implemented the ELF functionality in a new source file, with an
eye toward eliminating the a.out code entirely at some point in
the future.

26 years agoSync with sys/i386/conf/files.i386 revision 1.201.
KATO Takenori [Sat, 5 Sep 1998 02:36:18 +0000 (02:36 +0000)]
Sync with sys/i386/conf/files.i386 revision 1.201.

26 years agoSync with sys/i386/conf/Makefile.i386 revision 1.121.
KATO Takenori [Sat, 5 Sep 1998 02:35:21 +0000 (02:35 +0000)]
Sync with sys/i386/conf/Makefile.i386 revision 1.121.

26 years agoIncrease 'maxusers' to 32.
KATO Takenori [Sat, 5 Sep 1998 02:20:52 +0000 (02:20 +0000)]
Increase 'maxusers' to 32.

26 years agoWith the change to the 4.4Lite2 mount interface we don't need special
David E. O'Brien [Sat, 5 Sep 1998 00:58:37 +0000 (00:58 +0000)]
With the change to the 4.4Lite2 mount interface we don't need special
conf files.

26 years agoAdd hints file structures and defines.
John Polstra [Sat, 5 Sep 1998 00:53:53 +0000 (00:53 +0000)]
Add hints file structures and defines.

26 years agoEscape the \ characters so that the intended beep at the end of the
Mike Smith [Sat, 5 Sep 1998 00:36:18 +0000 (00:36 +0000)]
Escape the \ characters so that the intended beep at the end of the
boot.help message is actually appended correctly.

26 years agoAdd -lcrypt.
John Birrell [Sat, 5 Sep 1998 00:34:36 +0000 (00:34 +0000)]
Add -lcrypt.

26 years agoAdd -lcrypt when building kerberos.
John Birrell [Sat, 5 Sep 1998 00:32:27 +0000 (00:32 +0000)]
Add -lcrypt when building kerberos.

26 years agoActually *looking* at this file now, I see that a good deal more
Jordan K. Hubbard [Fri, 4 Sep 1998 23:51:40 +0000 (23:51 +0000)]
Actually *looking* at this file now, I see that a good deal more
can be eliminated. :)

26 years agoTurn tcp_extentions off by default. They sometimes cause pain.
Jordan K. Hubbard [Fri, 4 Sep 1998 23:37:24 +0000 (23:37 +0000)]
Turn tcp_extentions off by default.  They sometimes cause pain.

26 years agoMake irq forwarding truely functional.
Luoqi Chen [Fri, 4 Sep 1998 23:03:04 +0000 (23:03 +0000)]
Make irq forwarding truely functional.

26 years agoEnable rtld-elf for build on alpha too. We now have support for shared
John Birrell [Fri, 4 Sep 1998 22:55:17 +0000 (22:55 +0000)]
Enable rtld-elf for build on alpha too. We now have support for shared
libraries on alpha!

26 years agoUnbreak "make release".
Mark Murray [Fri, 4 Sep 1998 22:18:49 +0000 (22:18 +0000)]
Unbreak "make release".
Submitted by: Justin Gibbs

26 years agoYAMT (yet another missing target). It's a shame that these have to
John Birrell [Fri, 4 Sep 1998 21:19:37 +0000 (21:19 +0000)]
YAMT (yet another missing target). It's a shame that these have to
be visible to the user. Maybe `make release' should call the backend
directly.

26 years agoIncrease 'maxusers' to 32; with the number of people using GENERIC as
Mike Smith [Fri, 4 Sep 1998 19:55:51 +0000 (19:55 +0000)]
Increase 'maxusers' to 32; with the number of people using GENERIC as
their one-size-fits-all kernel, this should help reduce the "out of foo"
reports.

Reviewed by: jkh

26 years agoBring the docs more up-to-date. Document the 'custom' setup.
Andrzej Bialecki [Fri, 4 Sep 1998 19:47:24 +0000 (19:47 +0000)]
Bring the docs more up-to-date. Document the 'custom' setup.

26 years agoAdded special type of setup called 'custom'. It allows you to supply
Andrzej Bialecki [Fri, 4 Sep 1998 19:38:57 +0000 (19:38 +0000)]
Added special type of setup called 'custom'. It allows you to supply
your own config directory laid out exactly like the standard ones.

Suggested by: Richard Wackerbarth <rkw@dataplex.net>

26 years agoAdd alpha support.
Doug Rabson [Fri, 4 Sep 1998 19:03:57 +0000 (19:03 +0000)]
Add alpha support.

Submitted by: John Birrell <jb@cimlogic.com.au> (with extra hacks by me)
Obtained from: Probably NetBSD

26 years agoCosmetic changes to the PAGE_XXX macros to make them consistent with
Doug Rabson [Fri, 4 Sep 1998 18:49:35 +0000 (18:49 +0000)]
Cosmetic changes to the PAGE_XXX macros to make them consistent with
the other objects in vm.

26 years agoDon't cast potentially unaligned addresses to pointers to
Brian Somers [Fri, 4 Sep 1998 18:26:00 +0000 (18:26 +0000)]
Don't cast potentially unaligned addresses to pointers to
non-char types on non-i386 architectures.
On Alpha and Sparc we get a bus error if we do.

26 years agoDevice registration temporaly removed until major number allocation ok.
Nicolas Souchu [Fri, 4 Sep 1998 17:53:42 +0000 (17:53 +0000)]
Device registration temporaly removed until major number allocation ok.

26 years agoModify the xl_mediacheck() routine to also test for the case where the
Bill Paul [Fri, 4 Sep 1998 16:22:15 +0000 (16:22 +0000)]
Modify the xl_mediacheck() routine to also test for the case where the
XCVR value read from the EEPROM is completely wrong. I've had one report
of a 3c900 card that returns an xcvr value of 14, which is impossible
(the manual states that all vales above 8 are reserved). If the value
is out of the expe

Add PCI vendor ID for the 3c980-TX server adapter card, which apparently
also uses the cyclone chip. Graciously supplied Mats O Jansson
<maja@cntw.com>.

Also noted by Mats, the 10mpbs cyclone adapters should be named 3c900B,
not 3c905B. I haven't actually encountered a 10mbps only cyclone adapter
yet, nor anybody who has one, but this makes sense given the naming
scheme used for the older boomerang adapters.

26 years agoo Enable standby mode in apm utility. You need a newer kernel for this
Warner Losh [Fri, 4 Sep 1998 16:08:54 +0000 (16:08 +0000)]
o Enable standby mode in apm utility.  You need a newer kernel for this
  functionality to work (however, the kernel changes were committed
  to current at 1998/07/05 23:29:11 PDT).
o Enable setting of a timer for later wakeup.  Note, not all APM
  BIOSes support this for all modes (mine doesn't support waking up on
  a timer from standby mode).
o Print out the APM capabilities, as reported by the BIOS.
o Report wake on ring indication.  No way to enable/disable this, but
  it is reported.
o update man page with above
o The output of this command has changed, so if you have a script that
  parses its output, then you may need to rewrite it slightly.
o The code attempts to deal with older kernels than July 5, 1998, but
  that part of the code isn't well tested.

26 years agomake sbflush panic messages more descriptive
Andrey A. Chernov [Fri, 4 Sep 1998 13:13:18 +0000 (13:13 +0000)]
make sbflush panic messages more descriptive

26 years agoPAGE_WAKEUP -> vm_page_wakeup
Andrey A. Chernov [Fri, 4 Sep 1998 13:10:34 +0000 (13:10 +0000)]
PAGE_WAKEUP -> vm_page_wakeup

26 years agoFix uninitialized variable.
Kazutaka YOKOTA [Fri, 4 Sep 1998 10:15:48 +0000 (10:15 +0000)]
Fix uninitialized variable.

PR: bin/7799
Submitted by: Sheldon Hearn (axl@iafrica.com)

26 years agoAdd the distribute target that make release wants.
John Birrell [Fri, 4 Sep 1998 09:19:19 +0000 (09:19 +0000)]
Add the distribute target that make release wants.

26 years agoDon't utilize the pre-4.4Lite2 INITMOUNTNAMES compatibility cruft.
David E. O'Brien [Fri, 4 Sep 1998 08:55:29 +0000 (08:55 +0000)]
Don't utilize the pre-4.4Lite2 INITMOUNTNAMES compatibility cruft.
(bde will remove it soon)

26 years agoJust nuke the aout bootstrap stuff since we don't need it anymore.
Jordan K. Hubbard [Fri, 4 Sep 1998 08:22:59 +0000 (08:22 +0000)]
Just nuke the aout bootstrap stuff since we don't need it anymore.
Simplify accordingly for ELF.
Convinced by: jb

26 years agoSet OBJFORMAT=aout to affect the whole build - for now, it seems that too
Andrzej Bialecki [Fri, 4 Sep 1998 08:08:05 +0000 (08:08 +0000)]
Set OBJFORMAT=aout to affect the whole build - for now, it seems that too
many tools needed to build the floppies are dependent on a.out...

26 years agoCosmetic changes to the PAGE_XXX macros to make them consistent with
Doug Rabson [Fri, 4 Sep 1998 08:06:57 +0000 (08:06 +0000)]
Cosmetic changes to the PAGE_XXX macros to make them consistent with
the other objects in vm.