]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
29 years agoUse dsname() to get consistent names.
Bruce Evans [Wed, 22 Feb 1995 22:46:48 +0000 (22:46 +0000)]
Use dsname() to get consistent names.

29 years agoRemove the #error that temporarily stopped this from supporting slices and
Bruce Evans [Wed, 22 Feb 1995 22:06:20 +0000 (22:06 +0000)]
Remove the #error that temporarily stopped this from supporting slices and
labels.

Please test the slice/label features:
cd /dev; sh MAKEDEV svn0
cd /var/tmp; dd if=/dev/zero of=vnfile bs=8192 count=1024
vnconfig -c /dev/rvn0 /var/tmp/vnfile
fdisk /dev/rvn0  # invent a geometry, create one BSD partition
disklabel -r -w vn0 floppy  # a convenient (bogus) label
disklabel -e vn0  # edit label to match device
newfs /dev/rvn0a
mount /dev/vn0a /mnt
...
The steps after the fdisk could name the device vn0s1 instead of vn0.

29 years agoFix initialization of the copy of the label passed to setdisklabel().
Bruce Evans [Wed, 22 Feb 1995 21:51:53 +0000 (21:51 +0000)]
Fix initialization of the copy of the label passed to setdisklabel().

Fix the sign of the adjustment after writing a label.

Writing of labels should work now.

Merge adjust_label() into fixlabel().  Detect more errors and don't
write if there is an error.  Adjust sectors/unit and total sectors
to the numbers on the slice.

Add a function dsname() to print slice device names consistently, and
use it.

29 years agoAdd my thin 8x16 font with ISO-8859-1 characters.
Ollivier Robert [Wed, 22 Feb 1995 21:49:53 +0000 (21:49 +0000)]
Add my thin 8x16 font with ISO-8859-1 characters.
Sorry, no 8x8 or 8x14 font.

29 years agoDocument the mountdtab file in the man page.
Andreas Schulz [Wed, 22 Feb 1995 21:42:48 +0000 (21:42 +0000)]
Document the mountdtab file in the man page.

29 years agoAdd a prototype.
Bruce Evans [Wed, 22 Feb 1995 21:24:21 +0000 (21:24 +0000)]
Add a prototype.

29 years agoAdd ``Make find smarter about -fstype local''.
Rodney W. Grimes [Wed, 22 Feb 1995 20:28:07 +0000 (20:28 +0000)]
Add ``Make find smarter about -fstype local''.

29 years agoAdded v_interrupt_free_min to cnt struct.
David Greenman [Wed, 22 Feb 1995 18:06:49 +0000 (18:06 +0000)]
Added v_interrupt_free_min to cnt struct.

29 years agoCleanup for compatibility with another OS.
Stefan Eßer [Wed, 22 Feb 1995 18:04:56 +0000 (18:04 +0000)]
Cleanup for compatibility with another OS.

Submitted by: wolf

29 years agoMake good on my promise to finally clean up the config clobbering.
Jordan K. Hubbard [Wed, 22 Feb 1995 15:37:32 +0000 (15:37 +0000)]
Make good on my promise to finally clean up the config clobbering.
If you invoke config with the `-n' flag or have NO_CONFIG_CLOBBER in
your environment, config will behave the same way it used to.  This is
now _documented_ as well.  Rip out all the CONFIG_DONT_CLOBBER cruft;
some of it wasn't even correct anyway.

29 years agoNew PCI attach code:
Stefan Eßer [Wed, 22 Feb 1995 14:17:15 +0000 (14:17 +0000)]
New PCI attach code:

PCI BIOS mappings are retained, except if option PCI_REMAP
is specified in the kernel config file.
There is now a list of attach addresses, and the first
address that seems to make some device registers appear
is chosen.

Reviewed by: se
Submitted by: wolf

29 years agoAdded six new thin fonts (great on laptops)
Søren Schmidt [Wed, 22 Feb 1995 14:03:16 +0000 (14:03 +0000)]
Added six new thin fonts (great on laptops)
Cleaned up Makefile

29 years agoNext syscons update (given up on numbering :)
Søren Schmidt [Wed, 22 Feb 1995 13:48:07 +0000 (13:48 +0000)]
Next syscons update (given up on numbering :)

Removed screensavers from syscons, they are now LKM's. This makes it
possible to do some really "interesting" screensavers...
Fixed bug that sometimes caused garbage to appear when leaving
"scroll-lock" history.
Reformattet indentation, it got too deep for a normal 80 pos screen.
Split up in syscons.c & syscons.h for use with the saver-lkm's.
Temporarily removed -s option from vidcontrol, savers should now
be loaded with modload.

29 years agoAdd a man page for the Sony CDROM driver. Other driver authors, you might
Jordan K. Hubbard [Wed, 22 Feb 1995 13:30:39 +0000 (13:30 +0000)]
Add a man page for the Sony CDROM driver.  Other driver authors, you might
do a quick check here..

29 years agoChanged return value from vnode_pager_addr to be in DEV_BSIZE units so
David Greenman [Wed, 22 Feb 1995 10:34:34 +0000 (10:34 +0000)]
Changed return value from vnode_pager_addr to be in DEV_BSIZE units so
that 9 bits aren't lost in the conversion. Changed all callers to expect
this. This allows paging on large (>2GB) filesystems.

Submitted by: John Dyson

29 years agovm_page.c:
David Greenman [Wed, 22 Feb 1995 10:27:24 +0000 (10:27 +0000)]
vm_page.c:
Use request==VM_ALLOC_NORMAL rather than object!=kmem_object in deciding
if the caller is "important" in vm_page_alloc(). Also established a new
low threshold for non-interrupt allocations via cnt.v_interrupt_free_min.

vm_pageout.c:
Various algorithmic cleanup. Some calculations simplified. Initialize
cnt.v_interrupt_free_min to 2 pages.

Submitted by: John Dyson

29 years agoJust return in the case of a page not on any queue in vm_page_unqueue().
David Greenman [Wed, 22 Feb 1995 10:16:21 +0000 (10:16 +0000)]
Just return in the case of a page not on any queue in vm_page_unqueue().
Return VM_PAGE_BITS_ALL even if size > PAGE_SIZE in vm_page_bits().

Submitted by: John Dyson

29 years agoRemoved object locking code (it was a left over from an abortion that
David Greenman [Wed, 22 Feb 1995 10:06:43 +0000 (10:06 +0000)]
Removed object locking code (it was a left over from an abortion that
was done a month or so ago).

Submitted by: John Dyson

29 years agoRemoved bogus copy object collapse check (the idea is right, but the
David Greenman [Wed, 22 Feb 1995 10:00:16 +0000 (10:00 +0000)]
Removed bogus copy object collapse check (the idea is right, but the
spcific check was bogus).
Removed old copy of vm_object_page_clean and took out the #if 1 around
the remaining one.

Submitted by: John Dyson

29 years agovfs_cluster.c:
David Greenman [Wed, 22 Feb 1995 09:39:22 +0000 (09:39 +0000)]
vfs_cluster.c:
Various more tweaks from John Dyson to improve read ahead calculations.

vfs_subr.c:
Only wakeup if numoutput is 0 in vwakeup().

Submitted by: John Dyson

29 years agoAdded some code to make sure that buffers associated with directories and
David Greenman [Wed, 22 Feb 1995 09:30:13 +0000 (09:30 +0000)]
Added some code to make sure that buffers associated with directories and
metadata aren't thrashed by regular file I/O.
Added mechanism to limit the amount of outstanding I/O on a given vnode.
Pagedaemon wakeup policy changed to skew priority a little in favor of
file caching.
Slight code reorganization to improve clarity.
Added a few more comments.

Submitted by: John Dyson

29 years agoOnly do object paging_in_progress wakeups if someone is waiting on this
David Greenman [Wed, 22 Feb 1995 09:16:07 +0000 (09:16 +0000)]
Only do object paging_in_progress wakeups if someone is waiting on this
condition.
Added some comments.

Submitted by: John Dyson

29 years agoOnly do object paging_in_progress wakeups if someone is waiting on this
David Greenman [Wed, 22 Feb 1995 09:15:35 +0000 (09:15 +0000)]
Only do object paging_in_progress wakeups if someone is waiting on this
condition.

Submitted by: John Dyson

29 years agoRewrote MAP_PRIVATE case of vm_mmap() - all of the COW portion of this
David Greenman [Wed, 22 Feb 1995 08:40:54 +0000 (08:40 +0000)]
Rewrote MAP_PRIVATE case of vm_mmap() - all of the COW portion of this
routine was highly convoluted.

Submitted by: John Dyson

29 years agoFix benign type mismatch.
Bruce Evans [Wed, 22 Feb 1995 07:23:26 +0000 (07:23 +0000)]
Fix benign type mismatch.

29 years agoThe 'set ifilter'/'set ofilter' commands accept a syntax containing
Atsushi Murai [Wed, 22 Feb 1995 06:44:03 +0000 (06:44 +0000)]
The 'set ifilter'/'set ofilter' commands accept a syntax containing
IP addresses and/or protocol+port, but in the case where both are
supplied, it happily accepts the command but ignores the proto+port
It also attempts to handle the case where the second IP address
is omitted, but this doesn't work.

Reviewed by:  amurai@spec.co.jp
Submitted by: Andrew.Gordon@net-tel.co.uk

29 years agoadd /usr/local/share, needed by some ports
Andrey A. Chernov [Wed, 22 Feb 1995 04:34:19 +0000 (04:34 +0000)]
add /usr/local/share, needed by some ports

29 years agoAdded entry from Jim Bryant
Poul-Henning Kamp [Wed, 22 Feb 1995 03:49:46 +0000 (03:49 +0000)]
Added entry from Jim Bryant

29 years agoIncrease retry count while reading toc, old one
Andrey A. Chernov [Wed, 22 Feb 1995 02:12:10 +0000 (02:12 +0000)]
Increase retry count while reading toc, old one
isn't enough for spin up

29 years agoUse CDPLAY environment variable to not type argument each time
Andrey A. Chernov [Wed, 22 Feb 1995 01:45:03 +0000 (01:45 +0000)]
Use CDPLAY environment variable to not type argument each time
Submitted by: hvd@terry.ping.dk

29 years agoAdd a define for the SCSI_QUEUE_FULL sense message
Justin T. Gibbs [Wed, 22 Feb 1995 01:44:21 +0000 (01:44 +0000)]
Add a define for the SCSI_QUEUE_FULL sense message

29 years agoAdd tagged queueing support. Right now, we only allow two tagged
Justin T. Gibbs [Wed, 22 Feb 1995 01:43:25 +0000 (01:43 +0000)]
Add tagged queueing support.  Right now, we only allow two tagged
commands per target.  I could have followed the route of the ncr driver
and gone to great lengths to get the SCSI subsystem to support more, but
I think I'll use the time saved to help Julian and Peter make tagged
queuing a better handled generic feature.  This also includes some comment
and enum clean up and a possible fix for the hanging PCI controllers.

29 years agoAdd tagged queueing support to the aic7xxx sequencer code.
Justin T. Gibbs [Wed, 22 Feb 1995 01:37:52 +0000 (01:37 +0000)]
Add tagged queueing support to the aic7xxx sequencer code.

29 years agomixer added
Andrey A. Chernov [Wed, 22 Feb 1995 01:37:47 +0000 (01:37 +0000)]
mixer added

29 years agoMove mixer to usr.sbin
Andrey A. Chernov [Wed, 22 Feb 1995 01:36:12 +0000 (01:36 +0000)]
Move mixer to usr.sbin

29 years agoDo not close cd after eject, close on each ioctl error instead.
Andrey A. Chernov [Wed, 22 Feb 1995 01:19:06 +0000 (01:19 +0000)]
Do not close cd after eject, close on each ioctl error instead.
Do not open cd on help and quit commands.

29 years agoImplement tray closing and tray locking
Andrey A. Chernov [Wed, 22 Feb 1995 01:11:36 +0000 (01:11 +0000)]
Implement tray closing and tray locking
Fix first open fails bug

29 years agoObtained from: memories of 1.1.5
Bruce Evans [Tue, 21 Feb 1995 18:41:30 +0000 (18:41 +0000)]
Obtained from: memories of 1.1.5

Fix the sign of the timezone offset again.

29 years agoRemove FAT_CURSOR - it's deprecated.
Jordan K. Hubbard [Tue, 21 Feb 1995 13:42:57 +0000 (13:42 +0000)]
Remove FAT_CURSOR - it's deprecated.
Submitted by: "Philippe Charnier" <charnier@lirmm.fr>

29 years agoPanic if u_map allocation fails.
David Greenman [Tue, 21 Feb 1995 12:06:06 +0000 (12:06 +0000)]
Panic if u_map allocation fails.

29 years agoFix bogus Makefile which resulted in incompatible shared and static
Andras Olah [Tue, 21 Feb 1995 10:46:34 +0000 (10:46 +0000)]
Fix bogus Makefile which resulted in incompatible shared and static
libs.

Reviewed by: Michael Reifenberger <root@rz-wb.fh-sw.de>,
roberto@blaise.ibp.fr (Ollivier ROBERT)

29 years agoAdd sgsc to the clean/cleandir targets.
Andreas Schulz [Tue, 21 Feb 1995 10:30:41 +0000 (10:30 +0000)]
Add sgsc to the clean/cleandir targets.

29 years agoDiscard junk from partition tables when reading the tables for applications
Bruce Evans [Tue, 21 Feb 1995 08:38:24 +0000 (08:38 +0000)]
Discard junk from partition tables when reading the tables for applications
the same as when initializing the in-core copies.  Adjust checksums in
labels after adjusting labels.  This finishes fudging the on-disk label to
make it coherent with the in-core label.

Handle EIO during initialization better.

Initialize the compatibility slice to the whole disk If there are no real
slices.

Don't warn about adjusting offsets in the label to make the 'c' partition
start at 0.  The 'c' offset is now always absolute on-disk and 0 in-core
so an adjustment is usually required.

Don't confuse LABEL_PART with RAW_PART so much.

Check for partitions being within slices differently.

29 years agoDon't generate any errors here other than EIO. An invalid partition
Bruce Evans [Tue, 21 Feb 1995 08:37:09 +0000 (08:37 +0000)]
Don't generate any errors here other than EIO.  An invalid partition
table is not an error, it just causes a smaller slice table.

29 years agoSpeed md5 up around 30% by shorting out a couple of cumbersome
Poul-Henning Kamp [Tue, 21 Feb 1995 06:01:49 +0000 (06:01 +0000)]
Speed md5 up around 30% by shorting out a couple of cumbersome
memcpy equivalent functions.

29 years agoadd "fib"
Poul-Henning Kamp [Tue, 21 Feb 1995 04:41:03 +0000 (04:41 +0000)]
add "fib"

Reviewed by: phk
Submitted by: Mark Murray <mark@grondar.za>
Obtained from:

29 years agoNot sure why this one was -lcompat since it doesn't reference anything in
Garrett Wollman [Tue, 21 Feb 1995 04:35:43 +0000 (04:35 +0000)]
Not sure why this one was -lcompat since it doesn't reference anything in
it.

29 years agoUse POSIX.2 regular expressions, kill libcompat.
Garrett Wollman [Tue, 21 Feb 1995 04:32:55 +0000 (04:32 +0000)]
Use POSIX.2 regular expressions, kill libcompat.

29 years agoTotally get rid of the snic driver.
Jordan K. Hubbard [Tue, 21 Feb 1995 04:26:35 +0000 (04:26 +0000)]
Totally get rid of the snic driver.
Submitted by: davidg

29 years agoUse tcflush(), eliminate libcompat.
Garrett Wollman [Tue, 21 Feb 1995 04:15:15 +0000 (04:15 +0000)]
Use tcflush(), eliminate libcompat.

29 years agoUse POSIX.2 regular expressions, get rid of libcompat.
Garrett Wollman [Tue, 21 Feb 1995 04:05:20 +0000 (04:05 +0000)]
Use POSIX.2 regular expressions, get rid of libcompat.

29 years agomore now uses POSIX regular expressions and no longer needs libcompat.
Garrett Wollman [Tue, 21 Feb 1995 03:46:48 +0000 (03:46 +0000)]
more now uses POSIX regular expressions and no longer needs libcompat.

29 years agoBe more exclusive about cvs-locks.
Poul-Henning Kamp [Tue, 21 Feb 1995 02:09:47 +0000 (02:09 +0000)]
Be more exclusive about cvs-locks.

29 years agovm_extern.h: removed vm_allocate_with_pager.
David Greenman [Tue, 21 Feb 1995 01:32:47 +0000 (01:32 +0000)]
vm_extern.h: removed vm_allocate_with_pager.
Removed vm_user.c...it's now completely deprecated.

29 years agoRemoved vm_user.c.
David Greenman [Tue, 21 Feb 1995 01:31:14 +0000 (01:31 +0000)]
Removed vm_user.c.

29 years agoDeprecated remaining use of vm_deallocate. Deprecated vm_allocate_with_
David Greenman [Tue, 21 Feb 1995 01:22:48 +0000 (01:22 +0000)]
Deprecated remaining use of vm_deallocate. Deprecated vm_allocate_with_
pager(). Almost completely rewrote vm_mmap(); when John gets done with
the bottom half, it will be a complete rewrite. Deprecated most use of
vm_object_setpager(). Removed side effect of setting object persist
in vm_object_enter and moved this into the pager(s). A few other
cosmetic changes.

29 years agoSet page alloced for map entries as valid.
David Greenman [Tue, 21 Feb 1995 01:13:05 +0000 (01:13 +0000)]
Set page alloced for map entries as valid.

29 years agoAdded missing extern declaration for 'maxprocperuid'.
David Greenman [Tue, 21 Feb 1995 00:37:31 +0000 (00:37 +0000)]
Added missing extern declaration for 'maxprocperuid'.

29 years agoRemoved vm_allocate(), vm_deallocate(), and vm_protect() functions. The
David Greenman [Mon, 20 Feb 1995 23:58:10 +0000 (23:58 +0000)]
Removed vm_allocate(), vm_deallocate(), and vm_protect() functions. The
only function remaining in this file is vm_allocate_with_pager(), and this
will be going RSN. The file will be removed when this happens.

29 years agoUse of vm_allocate and vm_deallocate has been deprecated.
David Greenman [Mon, 20 Feb 1995 23:52:39 +0000 (23:52 +0000)]
Use of vm_allocate and vm_deallocate has been deprecated.

29 years agoMoved ACT_MAX, ACT_ADVANCE, and ACT_DECLINE to vm_page.h.
David Greenman [Mon, 20 Feb 1995 23:35:45 +0000 (23:35 +0000)]
Moved ACT_MAX, ACT_ADVANCE, and ACT_DECLINE to vm_page.h.

29 years agoUse of vm_allocate() and vm_deallocate() has been deprecated.
David Greenman [Mon, 20 Feb 1995 22:23:31 +0000 (22:23 +0000)]
Use of vm_allocate() and vm_deallocate() has been deprecated.

29 years agoFix restoring cpl in the swi_ast_phantom case. The system sometimes
Bruce Evans [Mon, 20 Feb 1995 20:35:24 +0000 (20:35 +0000)]
Fix restoring cpl in the swi_ast_phantom case.  The system sometimes
returned to user mode without enabling ASTs.  The problem fixed itself
at the next syscall or non-FPU trap, if any.  It hung the system for
a test process that masked SIGFPE's and divided by zero.  The faulting
division was returned to endlessly and this gave plently of opportunities
for the swi_ast_phantom case to be reached; after it was reached the
system hung because the ASTs for preemption and SIGINT handling were
disabled.

29 years agoImplement maxprocperuid and maxfilesperproc. They are tunable
Guido van Rooij [Mon, 20 Feb 1995 19:42:42 +0000 (19:42 +0000)]
Implement maxprocperuid and maxfilesperproc. They are tunable
via sysctl(8). The initial value of maxprocperuid is maxproc-1,
that of maxfilesperproc is maxfiles (untill maxfile will disappear)

Now it is at least possible to prohibit one user opening maxfiles

-Guido

Submitted by:
Obtained from:

29 years agoUse absolute path for the "ctm" program.
Poul-Henning Kamp [Mon, 20 Feb 1995 19:07:32 +0000 (19:07 +0000)]
Use absolute path for the "ctm" program.

29 years agoFixed non-conflict errors that occur when changes made to the local files are
Nate Williams [Mon, 20 Feb 1995 18:47:18 +0000 (18:47 +0000)]
Fixed non-conflict errors that occur when changes made to the local files are
the same as the changes made in the repository.  This is often seen by
people with remote CVS trees that have applied their local patches to the
master site.  a 'cvs update' will show bogus conflicts.

Obtained from: CVS mailing list, Stig<stig@inse.com>

In diffutils 2.6 and 2.7, diff3 -A complains about identical overlapping
changes.  They're different from the ancestor but not from each other...
Why bother?  The patch below fixes this nonsense and preserves [B]ackwards
compatiblity with the -B flag (also --show-bogus-conflicts).

    Party on...
    Stig

29 years agoMake libcompat a static only library.
Nate Williams [Mon, 20 Feb 1995 18:19:50 +0000 (18:19 +0000)]
Make libcompat a static only library.

Since functions will come and go from libcompat as they are deprecated
it makes no sense to build a shared library out of it as it will change.

Based on freedback from Terry and Jonas on the mailing lists.

29 years agovm_inherit function has been deprecated.
David Greenman [Mon, 20 Feb 1995 18:08:18 +0000 (18:08 +0000)]
vm_inherit function has been deprecated.

29 years agoStop using vm_allocate and vm_deallocate.
David Greenman [Mon, 20 Feb 1995 17:36:36 +0000 (17:36 +0000)]
Stop using vm_allocate and vm_deallocate.

29 years agoVM for the kernel stack and page tables doesn't need to be explicitly
David Greenman [Mon, 20 Feb 1995 17:26:21 +0000 (17:26 +0000)]
VM for the kernel stack and page tables doesn't need to be explicitly
deallocated as it isn't inherited across the fork.
Use vm_map_find not vm_allocate.

Submitted by: John Dyson

29 years agoAdd Ollivier and Satoshi to list of contributors.
Jordan K. Hubbard [Mon, 20 Feb 1995 16:24:30 +0000 (16:24 +0000)]
Add Ollivier and Satoshi to list of contributors.

29 years agoMake sure process isn't swapped when messing with it.
David Greenman [Mon, 20 Feb 1995 15:53:33 +0000 (15:53 +0000)]
Make sure process isn't swapped when messing with it.
Added missing newline to log() call.

29 years agoAdded missing newlines to calls to log().
David Greenman [Mon, 20 Feb 1995 15:48:46 +0000 (15:48 +0000)]
Added missing newlines to calls to log().

29 years agoPanic if object is deallocated too many times.
David Greenman [Mon, 20 Feb 1995 14:21:58 +0000 (14:21 +0000)]
Panic if object is deallocated too many times.
Slight change to reverse collapsing so that vm_object_deallocate doesn't
have to be called recursively.
Removed half of a previous fix - the renamed page during a collapse doesn't
need to be marked dirty because the pager backing store pointers are copied
- thus preserving the page's data. This assumes that pages without backing
store are always dirty (except perhaps for when they are first zeroed, but
this doesn't matter).
Switch order of two lines of code so that the correct pager is removed
from the hash list. The previous code bogusly passed a NULL pointer to
vm_object_remove(). The call to vm_object_remove() should be unnecessary
if named anonymous objects were being dealt with correctly. They are
currently marked as OBJ_INTERNAL, which really screws up things (such as
this).

29 years agoDon't allow act_count to exceed ACT_MAX when bumping it up.
David Greenman [Mon, 20 Feb 1995 14:00:50 +0000 (14:00 +0000)]
Don't allow act_count to exceed ACT_MAX when bumping it up.
Small optimization to vm_page_bits().

Submitted by: John Dyson

29 years agoFully initialize pages returned via vm_page_alloc_contig() so that the
David Greenman [Mon, 20 Feb 1995 13:58:14 +0000 (13:58 +0000)]
Fully initialize pages returned via vm_page_alloc_contig() so that the
memory can be later freed.

29 years agoAdd Gary J. for all his fine work with gdb and other things.
Jordan K. Hubbard [Mon, 20 Feb 1995 12:45:09 +0000 (12:45 +0000)]
Add Gary J. for all his fine work with gdb and other things.

29 years agoThe new and improved mkCTM. Learning from this experience it has been
Poul-Henning Kamp [Mon, 20 Feb 1995 03:03:02 +0000 (03:03 +0000)]
The new and improved mkCTM.  Learning from this experience it has been
improved on a couple of accounts.  Amongst these are "damage control"
more than 100 files removed and it will bail out...

29 years agoClean this file up so it is readable...
Poul-Henning Kamp [Mon, 20 Feb 1995 00:48:50 +0000 (00:48 +0000)]
Clean this file up so it is readable...

29 years agoDon't flush the message buffer when it fills up.
Bruce Evans [Sun, 19 Feb 1995 15:02:26 +0000 (15:02 +0000)]
Don't flush the message buffer when it fills up.

29 years agoRemove the SCSI idsn line - it's permanantly dead.
Jordan K. Hubbard [Sun, 19 Feb 1995 14:07:28 +0000 (14:07 +0000)]
Remove the SCSI idsn line - it's permanantly dead.

29 years agoremove gnu/scsi/nic5000.c - it's deprecated.
Jordan K. Hubbard [Sun, 19 Feb 1995 14:06:29 +0000 (14:06 +0000)]
remove gnu/scsi/nic5000.c - it's deprecated.

29 years agoBruces fixes for large filesystems
Gary Palmer [Sun, 19 Feb 1995 10:36:17 +0000 (10:36 +0000)]
Bruces fixes for large filesystems

Submitted by: Bruce Evans <bde@zeta.org.au>

29 years agoWhoops! It wasn't totally winning. I left out a couple of :'s. Fixed.
Jordan K. Hubbard [Sun, 19 Feb 1995 09:52:52 +0000 (09:52 +0000)]
Whoops!  It wasn't totally winning.  I left out a couple of :'s.  Fixed.

29 years agoA much more winning default dir file.
Jordan K. Hubbard [Sun, 19 Feb 1995 09:50:27 +0000 (09:50 +0000)]
A much more winning default dir file.

29 years agoChange DIR to VOLUME so that this installs in the correct location.
Rodney W. Grimes [Sun, 19 Feb 1995 09:11:30 +0000 (09:11 +0000)]
Change DIR to VOLUME so that this installs in the correct location.

29 years ago1. Get all the nodes in this texinfo document happily pointing in the proper
Jordan K. Hubbard [Sun, 19 Feb 1995 09:11:07 +0000 (09:11 +0000)]
1. Get all the nodes in this texinfo document happily pointing in the proper
directions.

2. Drop in a Makefile that properly installs the info files for this.

29 years agoSome recent small changes. Probably about time I start writing this in C.
Poul-Henning Kamp [Sun, 19 Feb 1995 04:51:41 +0000 (04:51 +0000)]
Some recent small changes.  Probably about time I start writing this in C.

29 years agoTruncate the pte address to a page boundry. This probably won't fix the
David Greenman [Sun, 19 Feb 1995 02:50:31 +0000 (02:50 +0000)]
Truncate the pte address to a page boundry. This probably won't fix the
panic, but at least it's more correct.

29 years agoKeep absolute offsets in on-disk labels for backwards compatibility. This
Bruce Evans [Sat, 18 Feb 1995 22:10:44 +0000 (22:10 +0000)]
Keep absolute offsets in on-disk labels for backwards compatibility.  This
requires complications to adjust the offsets to relative when a block
containing the label is read and back to absolute when such a block is
written.  The adjustment is not made on the whole disk slice.

Don't allow setting the offset of partition C to nonzero in in-core labels.
This will cause some (nonstandard) disktab entries to fail.  They will
need to be changed to have relative offsets (and no partitions outside
of the slice).

Don't write protect the (nonexistent) label on the whole disk slice.

Writing labels and bootstraps should work right now (except if there is
no DOSpartition table).

29 years agoWrite setconf() in KNF, in K&R C, and in Standard C.Standard C.
Bruce Evans [Sat, 18 Feb 1995 21:32:19 +0000 (21:32 +0000)]
Write setconf() in KNF, in K&R C, and in Standard C.Standard C.

29 years agoNew field b_biodone_chain to support nested b_iodone's.
Bruce Evans [Sat, 18 Feb 1995 21:12:33 +0000 (21:12 +0000)]
New field b_biodone_chain to support nested b_iodone's.

29 years agoSubmitted by: Ollivier ROBERT roberto@keltia.frmug.fr.net
Andreas Schulz [Sat, 18 Feb 1995 20:44:55 +0000 (20:44 +0000)]
Submitted by: Ollivier ROBERT roberto@keltia.frmug.fr.net
There is a typo in the include file specification in the  man page  of
time(1).

29 years agoDo away with 'options SWAP_GENERIC' once and for all: I get ill
Bill Paul [Sat, 18 Feb 1995 18:05:35 +0000 (18:05 +0000)]
Do away with 'options SWAP_GENERIC' once and for all: I get ill
just thinking about it.

Two changes need to be made to allow 'config kernel swap generic' to
work properly without requiring any compile-time flags:

/usr/src/usr.sbin/config/mkswapconf.c: we need to define a dummy stub
for the setconf() function to replace the one in swapgeneric.c that
isn't available in non-generic configurations.

/usr/src/sys/i386/i386/autoconf.c: the -a boot flag causes setroot()
to be skipped and lets setconf() prompt the user for a root device.
If you skip setroot() in a non-generic kernel, you could get severely
hosed. To avoid this, we silently ignore the -a flag if rootdev != NODEV.
(rootdev is always initialized to NODEV in swapgeneric.c, so if
we find that rootdev is something other than NODEV, we know we're
not using a generic configuration.)

29 years agoRemove bogus reference to snp.4. Somebody kinda screwed up their move! :)
Jordan K. Hubbard [Sat, 18 Feb 1995 17:10:40 +0000 (17:10 +0000)]
Remove bogus reference to snp.4.  Somebody kinda screwed up their move! :)

29 years agoipfirewall.4 is obviously not here anymore! Adjust the Makefile.
Jordan K. Hubbard [Sat, 18 Feb 1995 16:36:23 +0000 (16:36 +0000)]
ipfirewall.4 is obviously not here anymore!  Adjust the Makefile.

29 years agoMinor optimization.
Andrey A. Chernov [Sat, 18 Feb 1995 11:36:33 +0000 (11:36 +0000)]
Minor optimization.

29 years agoAdd Nate back as promised before, but I will also note that he's on
Jordan K. Hubbard [Sat, 18 Feb 1995 09:31:16 +0000 (09:31 +0000)]
Add Nate back as promised before, but I will also note that he's on
probation..  Another vanishing act and I won't hesitate to nuke this back
out again.  He hasn't been heard from very much since he was over here to
visit! :-)

29 years agoCorrect the grammar for the label of the number of users.
Scott Mace [Sat, 18 Feb 1995 08:24:16 +0000 (08:24 +0000)]
Correct the grammar for the label of the number of users.
It is plural when zero and >1 users are logged in.

29 years ago1) Added protection against collapsing OBJ_DEAD objects.
David Greenman [Sat, 18 Feb 1995 06:48:33 +0000 (06:48 +0000)]
1) Added protection against collapsing OBJ_DEAD objects.
2) bump reference counts by 2 instead of 1 so that an object deallocate
   doesn't try to recursively collapse the object.
3) mark pages renamed during the collapse as dirty so that their contents
   are preserved.

Submitted by: John and me.