]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoThis commit was generated by cvs2svn to compensate for changes in r60484,
David E. O'Brien [Fri, 12 May 2000 23:15:20 +0000 (23:15 +0000)]
This commit was generated by cvs2svn to compensate for changes in r60484,
which included commits to RCS files with non-trunk default branches.

24 years agoImport of Binutils 2.10 snapshot.
David E. O'Brien [Fri, 12 May 2000 23:15:20 +0000 (23:15 +0000)]
Import of Binutils 2.10 snapshot.

24 years agoUse a alternate side 2.9.1 source tree while upgrading Binutils.
David E. O'Brien [Fri, 12 May 2000 22:55:23 +0000 (22:55 +0000)]
Use a alternate side 2.9.1 source tree while upgrading Binutils.

24 years agoReactivate the FICL hooks to make it be compiled in, but also initialize
Peter Wemm [Fri, 12 May 2000 22:45:16 +0000 (22:45 +0000)]
Reactivate the FICL hooks to make it be compiled in, but also initialize
FICL.  bootforth is now live on the Alpha!

**BEWARE** - you *MUST* build and install a current libstand or you will
most likely get zfree() panics at loader startup.

We should now be able to set up the loader.conf stuff on the Alpha too.

24 years agoFix the real problem that broke the Alpha loader this last week. It
Peter Wemm [Fri, 12 May 2000 22:43:20 +0000 (22:43 +0000)]
Fix the real problem that broke the Alpha loader this last week.  It
was not the fault of the module code, nor FICL.  The malloc code requires
sbrk() to return addresses that were at least 16 byte aligned.  If the
Alpha loader happened to be 8 byte but not 16 byte aligned in length, then
you would get a zfree() panic at startup.

Incidently, this affected the i386 loader as well, and explains why
the static heap changed things and why jlemon had trouble when the bss
was not ending at a multiple of 8 bytes.

My fix is to 16 byte align it on all arches, even though the x86 version
only required 8 byte alignment (struct MemNode is smaller there).  We could
page align it if we wanted to be paranoid, but it isn't presently necessary.

24 years agoStop libficl from using floating point registers - this was cauing
Peter Wemm [Fri, 12 May 2000 21:51:20 +0000 (21:51 +0000)]
Stop libficl from using floating point registers - this was cauing
dictDelete() to fault early on the Alpha and was the original cause of
the Alpha ficl failures.

24 years agoWe compile expressly on the alpha with -mno-fp-regs - so do not use them
Peter Wemm [Fri, 12 May 2000 21:48:54 +0000 (21:48 +0000)]
We compile expressly on the alpha with -mno-fp-regs - so do not use them
explicitly here. This was the original cause of instruction faults on
the Alpha in loader. (not this code, but the same problem in libficl.a)

24 years agoNull commit:
Chris Costello [Fri, 12 May 2000 20:52:50 +0000 (20:52 +0000)]
Null commit:
  The log entry for the previous commit marked an item
  relating to filling in [mca]time data for socket structures
  as related to FIFOs.  The change is actually related
  to sockets.

24 years agoBackout previous commit to this file: it dies in buildworld environment.
Tim Vanderhoek [Fri, 12 May 2000 18:10:17 +0000 (18:10 +0000)]
Backout previous commit to this file: it dies in buildworld environment.
I probably forgot to put an ${.OBJDIR} somewhere.  I'll redo the change
later when I get a chance to test that thesis.

24 years agoThis driver attempts to cache mbufs, and re-use them as headers instead
Jonathan Lemon [Fri, 12 May 2000 16:48:59 +0000 (16:48 +0000)]
This driver attempts to cache mbufs, and re-use them as headers instead
of using the MGETHDR macro all the time.  When an mbuf is reused as a
header, initialize csum_flags to zero as well, so the delayed_checksum
call woks properly.

Debbugging work done by: jmas

24 years agoFix comment typo.
Jeroen Ruigrok van der Werven [Fri, 12 May 2000 16:06:49 +0000 (16:06 +0000)]
Fix comment typo.

Submitted by: nrahlstr

24 years agoCorrect duplicate use of word "been" in a sentence.
Paul Richards [Fri, 12 May 2000 15:31:54 +0000 (15:31 +0000)]
Correct duplicate use of word "been" in a sentence.

24 years agoUse bus_space stuff except where it needs high performance.
Yoshihiro Takahashi [Fri, 12 May 2000 12:38:25 +0000 (12:38 +0000)]
Use bus_space stuff except where it needs high performance.

24 years agoUse bus_space stuff except where it needs high performance.
Yoshihiro Takahashi [Fri, 12 May 2000 12:37:28 +0000 (12:37 +0000)]
Use bus_space stuff except where it needs high performance.

Reviewed by: bde

24 years agoFix typo
Alexey Zelkin [Fri, 12 May 2000 10:50:04 +0000 (10:50 +0000)]
Fix typo

Spoted by: bde

24 years agomdoc related fixes:
Alexey Zelkin [Fri, 12 May 2000 10:22:50 +0000 (10:22 +0000)]
mdoc related fixes:
. synchronize NAME and SYNOPSIS sections
. replace .Ev macros with .Dv / .Er / .Em macros  as mdoc(7)
  specification declare

24 years ago. fix .Dt macro argument
Alexey Zelkin [Fri, 12 May 2000 10:07:31 +0000 (10:07 +0000)]
. fix .Dt macro argument
. spell inet6_rthdr_reverse correctly

24 years agoadd MLINKS: pthread_testcancel(3) -> pthread_setcanceltype(3)
Alexey Zelkin [Fri, 12 May 2000 09:59:44 +0000 (09:59 +0000)]
add MLINKS: pthread_testcancel(3) -> pthread_setcanceltype(3)

24 years ago. insert forgotten by someone "Space Mode on" directive. It caused displaying
Alexey Zelkin [Fri, 12 May 2000 09:10:40 +0000 (09:10 +0000)]
. insert forgotten by someone "Space Mode on" directive. It caused displaying
  of all macros placed after FILES section w/o any space characters.
. slightly correct usage of .An macros.

24 years agoPlace $FreeBSD$ tag to correct location (bottom of the copyright)
Alexey Zelkin [Fri, 12 May 2000 08:37:08 +0000 (08:37 +0000)]
Place $FreeBSD$ tag to correct location (bottom of the copyright)

24 years agoFix driver source location path (/sys/i386/isa/fd.c -> /sys/isa/fd.c)
Alexey Zelkin [Fri, 12 May 2000 08:36:36 +0000 (08:36 +0000)]
Fix driver source location path (/sys/i386/isa/fd.c -> /sys/isa/fd.c)

24 years agoUse .An macro for declare Authors Name in AUTHORS section
Alexey Zelkin [Fri, 12 May 2000 08:32:56 +0000 (08:32 +0000)]
Use .An macro for declare Authors Name in AUTHORS section

24 years agoAdd MLINKS: scsi.4 -> SCSI.4, scsi.4 -> CAM.4
Alexey Zelkin [Fri, 12 May 2000 08:32:09 +0000 (08:32 +0000)]
Add MLINKS: scsi.4 -> SCSI.4, scsi.4 -> CAM.4

24 years ago. replace CONFIGURATION section name with SYNOPSYS as mdoc(7) style declare.
Alexey Zelkin [Fri, 12 May 2000 08:30:58 +0000 (08:30 +0000)]
. replace CONFIGURATION section name with SYNOPSYS as mdoc(7) style declare.
. remove redundant paragraph breaks.
. fix typo (.Sj -> .Sh)

24 years agoProperly mdoc'ify this manual page: remove all direct troff formating
Alexey Zelkin [Fri, 12 May 2000 08:28:01 +0000 (08:28 +0000)]
Properly mdoc'ify this manual page: remove all direct troff formating
directives, apply correct mdoc(7) Blocks and Lists formatting.

24 years agoAdd MLINKS: bktr.4 -> booktree.4
Alexey Zelkin [Fri, 12 May 2000 08:14:11 +0000 (08:14 +0000)]
Add MLINKS: bktr.4 -> booktree.4

24 years agoRemove osolete reference to eg(4), properly use mdoc macro `.Nm'
Alexey Zelkin [Fri, 12 May 2000 08:10:22 +0000 (08:10 +0000)]
Remove osolete reference to eg(4), properly use mdoc macro `.Nm'

24 years agoAdd MLINKS for all macros decribed in queue.3 manpage.
Alexey Zelkin [Fri, 12 May 2000 07:58:37 +0000 (07:58 +0000)]
Add MLINKS for all macros decribed in queue.3 manpage.

24 years agoFix the Alpha loader the rest of the way. If FICL is present in
Peter Wemm [Fri, 12 May 2000 07:47:47 +0000 (07:47 +0000)]
Fix the Alpha loader the rest of the way.  If FICL is present in
/boot/loader (even though it is 100% dormant in the Alpha version),
then the loader panics with a zfree error:Loading /boot/loader.test
  *** keyboard not plugged in...
  Console: SRM firmware console
  panic: zfree(0x2003cb58,4096): wild pointer
versus the exact same code but without FICL linked in:
  Loading /boot/loader
  Console: SRM firmware console
  VMS PAL rev: 0x1000600010114
  OSF PAL rev: 0x1000600020116
  Switch to OSF PAL code succeeded.

  FreeBSD/alpha SRM disk boot, Revision 0.1

This is almost certainly an alpha infrastructure bug, not a FICL
problem.  It's probably the same thing that made FICL fail for no
apparent reason on the Alpha.

24 years agolong != int on Alphas.
Boris Popov [Fri, 12 May 2000 05:16:37 +0000 (05:16 +0000)]
long != int on Alphas.

24 years agoImprove hack from previous commit to this file: exit if we get successive
Tim Vanderhoek [Fri, 12 May 2000 04:04:27 +0000 (04:04 +0000)]
Improve hack from previous commit to this file: exit if we get successive
EOFs from reading stderr (eg. not from argv[1]).

24 years agoUse termcap(5) function key sequences rather than hardcoding for syscons.
Tim Vanderhoek [Fri, 12 May 2000 04:02:13 +0000 (04:02 +0000)]
Use termcap(5) function key sequences rather than hardcoding for syscons.
Now page-up/down work from xterms.

24 years agoCreate magic variables that return termcap(5) strings for function keys.
Tim Vanderhoek [Fri, 12 May 2000 04:00:23 +0000 (04:00 +0000)]
Create magic variables that return termcap(5) strings for function keys.

24 years agoFix an uncommon bug that would cause us to stop accepting input if the
Tim Vanderhoek [Fri, 12 May 2000 03:53:38 +0000 (03:53 +0000)]
Fix an uncommon bug that would cause us to stop accepting input if the
user entered a command that filled exactly the remaining screen width.

24 years agoOnly main.c depends on defrc.h
Tim Vanderhoek [Fri, 12 May 2000 03:52:03 +0000 (03:52 +0000)]
Only main.c depends on defrc.h

24 years agoAdd support for the DVD ioctl interface.
Kenneth D. Merry [Fri, 12 May 2000 03:36:02 +0000 (03:36 +0000)]
Add support for the DVD ioctl interface.

24 years agoActually accept 'm' as a PAGER query response to mean 'more'.
Brian S. Dean [Fri, 12 May 2000 03:09:57 +0000 (03:09 +0000)]
Actually accept 'm' as a PAGER query response to mean 'more'.

24 years agoTurn on USB support for most USB devices. udbp is not turned on since
John Baldwin [Fri, 12 May 2000 03:05:35 +0000 (03:05 +0000)]
Turn on USB support for most USB devices.  udbp is not turned on since
NETGRAPH is not present in GENERIC at the moment.  Also, change some
settings to support USB installs:

- Add KBD_INSTALL_CDEV as an option to make /dev/kbd[01] actually work.
- Turn on keyboard probing in sc0.  The syscons driver will now use a
  flag documented in ukbd(4) but not in sc(4) that tells syscons to
  actively search for a keyboard device if none is found.  This allows
  USB keyboards to just be plugged in and instantly start working.
- Require the atkbd0 driver to actually probe to see if a keyboard is
  there.  This allows USB keyboards to be seen by sc0 if an AT keyboard
  isn't plugged into the computer.  This also means that you will no
  longer be able to plug an AT keyboard into a machine after it has
  booted a GENERIC kernel and use it.  AT keyboards aren't designed for
  this anyway.  USB keyboards are designed for this, and they work.

24 years agoGive mtree the ability to exclude files and directories from its traversal.
Garrett Wollman [Fri, 12 May 2000 03:03:00 +0000 (03:03 +0000)]
Give mtree the ability to exclude files and directories from its traversal.

PR: 16944

24 years agoAdd support for USB to sysinstall. This includes running usbd and
John Baldwin [Fri, 12 May 2000 03:01:17 +0000 (03:01 +0000)]
Add support for USB to sysinstall.  This includes running usbd and
setting 'usbd_enable' in rc.conf during nwe installs if USB is detected.
Also, since usbd already handles USB mice automatically, note that the
mouse setup section in sysinstall only applies to non-USB mice.

24 years agoAdd USB devices to /dev on the install floppy.
John Baldwin [Fri, 12 May 2000 02:58:59 +0000 (02:58 +0000)]
Add USB devices to /dev on the install floppy.
Add usbd.conf to /etc on the install floppy.

24 years agoAdd usbd and usbdevs to the install floppy in /stand.
John Baldwin [Fri, 12 May 2000 02:57:56 +0000 (02:57 +0000)]
Add usbd and usbdevs to the install floppy in /stand.

24 years agoAdd kbd1 to MAKEDEV's all) target so that we support using kbdcontrol
John Baldwin [Fri, 12 May 2000 02:56:31 +0000 (02:56 +0000)]
Add kbd1 to MAKEDEV's all) target so that we support using kbdcontrol
to switch keyboards out of the box.

24 years agoBetter fix for strftime/mktime bug, now more compatible since local timezone
Andrey A. Chernov [Thu, 11 May 2000 23:49:13 +0000 (23:49 +0000)]
Better fix for strftime/mktime bug, now more compatible since local timezone
name and offset restored for compatibility with standard perl

24 years agoFixed small bug introduced in my last commit.
Murray Stokely [Thu, 11 May 2000 23:43:12 +0000 (23:43 +0000)]
Fixed small bug introduced in my last commit.

Approved by: jkh

24 years agoTemporarily turn off the newreno flag until we can track down the known
Jayanth Vijayaraghavan [Thu, 11 May 2000 22:28:28 +0000 (22:28 +0000)]
Temporarily turn off the newreno flag until we can track down the known
data corruption problem.

24 years agoDocument the `new' fdesc--now that it belongs only on /dev/fd, remove all
Chris Costello [Thu, 11 May 2000 22:20:23 +0000 (22:20 +0000)]
Document the `new' fdesc--now that it belongs only on /dev/fd, remove all
non-/dev/fd related documentation (such as that for stdin, stdout, tty).

Update the fstab line to refer to /dev/fd instead of /dev.

24 years agoAdapt fdesc to be mounted on /dev/fd and remove fd, stdin, stdout and
Chris Costello [Thu, 11 May 2000 22:10:51 +0000 (22:10 +0000)]
Adapt fdesc to be mounted on /dev/fd and remove fd, stdin, stdout and
stderr nodes.  More specific items of this patch:
  o Removed support for symbolic links, and the need for
    fdesc_readlink().
  o Put all the code from fdesc_attr() into fdesc_getattr() and removed
    fdesc_attr().  This also made it easier to properly give all nodes
    unique inode numbers.
  o The removal of all non-fd nodes allowed the removal of the fdesc_read(),
    fdesc_write(), and fdesc_ioctl() nodes, since we no longer have nodes
    that get special handling.
  o Correct the component name validity-checking in fdesc_lookup().  It
    previously detected the end of the string by checking for a terminating
    NUL, now it uses cnp->cn_namelen.
  o Handle kqueue files as FIFOs.  This is probably the closest file type
    to represent this type of file there is, and it is unfortunately not
    very representative of a kqueue.  Creation time is not supported by
    kqueue, so ctime, mtime and atime are all set to the current time when
    getattr() was called.
  o Also set st_[mca]time to the current time since there's no data in
    socket structures that can be used to fill this in (FIFOs).
  o Simplify fdesc_readdir() since it only has to report the numbered
    fd nodes.  Add `.' and `..' directory links as well.
  o Remove read bits from directories as they tend to confuse programs
    like tar(1).

Reviewed by: phk
Discussed with: bde (earlier on, not quite review)

24 years agoInclude the UID and GID values filled in by socreate() into socket->so_cred
Chris Costello [Thu, 11 May 2000 22:08:57 +0000 (22:08 +0000)]
Include the UID and GID values filled in by socreate() into socket->so_cred
for stat() calls.

Reviewed by: phk

24 years agoInclude UID and GID information for stat() calls using the values filled
Chris Costello [Thu, 11 May 2000 22:08:20 +0000 (22:08 +0000)]
Include UID and GID information for stat() calls using the values filled
into the file descriptor data by falloc().

Reviewed by: phk

24 years agoOnly display the 'Disable PNP-OS BIOS option' message on the i386 arch.
John Baldwin [Thu, 11 May 2000 20:17:28 +0000 (20:17 +0000)]
Only display the 'Disable PNP-OS BIOS option' message on the i386 arch.
Alpha's don't have a PNP BIOS option.

Reviewed by: n_hibma

24 years agoSmall style fix '=' -> ' = '
Nick Sayer [Thu, 11 May 2000 20:15:16 +0000 (20:15 +0000)]
Small style fix '=' -> ' = '

24 years agoClaim support / explain white-box Alphas.
Wilko Bulte [Thu, 11 May 2000 18:56:45 +0000 (18:56 +0000)]
Claim support / explain white-box Alphas.

Obtained from: drew

24 years agoNewbusify resource allocation.
Duncan Barclay [Thu, 11 May 2000 18:55:38 +0000 (18:55 +0000)]
Newbusify resource allocation.

Have OLDCARD version of the remapping.

Remove BPF conditionals.

Remove ISA/3.x stuff and add -current stuff.

24 years agoRemove ISA/3.x stuff and add -current stuff
Duncan Barclay [Thu, 11 May 2000 18:53:50 +0000 (18:53 +0000)]
Remove ISA/3.x stuff and add -current stuff

Use device_printf.

24 years agoUse device_printf.
Duncan Barclay [Thu, 11 May 2000 18:53:10 +0000 (18:53 +0000)]
Use device_printf.

24 years agoFixed a syntax error (a stray semicolon caused by a bogus #define).
Bruce Evans [Thu, 11 May 2000 17:10:22 +0000 (17:10 +0000)]
Fixed a syntax error (a stray semicolon caused by a bogus #define).

24 years agoName the `struct bio *' parameter of devstat_end_transaction_bio().
Bruce Evans [Thu, 11 May 2000 17:07:24 +0000 (17:07 +0000)]
Name the `struct bio *' parameter of devstat_end_transaction_bio().

Removed the last vestige of devstat_end_transaction_buf().

24 years agoFix compatibility issue in sed. Do so by explicitely adding a
Nick Sayer [Thu, 11 May 2000 17:01:52 +0000 (17:01 +0000)]
Fix compatibility issue in sed. Do so by explicitely adding a
newline to the end of any -e argument.

PR: bin/18474

24 years agoFixed missing prototype for inet6_rthdr_reverse().
Bruce Evans [Thu, 11 May 2000 16:57:45 +0000 (16:57 +0000)]
Fixed missing prototype for inet6_rthdr_reverse().

24 years agoAbuse .Fn for macros better (don't hand-format the parameter string by
Bruce Evans [Thu, 11 May 2000 16:54:52 +0000 (16:54 +0000)]
Abuse .Fn for macros better (don't hand-format the parameter string by
quoting it all and adding commas).

Don't say that the expression in KASSERT() is an int.  It is a collection
of tokens forming a C expression that can be compared with 0.

24 years agoRemoved bogus quoting of single tokens in a .Fn invocation. This helps
Bruce Evans [Thu, 11 May 2000 16:39:33 +0000 (16:39 +0000)]
Removed bogus quoting of single tokens in a .Fn invocation.  This helps
my synopsis checker decide which .Fn's are for prototypes (most function
parameters require quoting because they have multiple tokens).

24 years agoFixed bitrot. The changes from buf to bio had not reached here.
Bruce Evans [Thu, 11 May 2000 16:25:03 +0000 (16:25 +0000)]
Fixed bitrot.  The changes from buf to bio had not reached here.

24 years agoAbuse .Ft and .Fn for macros in the usual way (name the macro parameters,
Bruce Evans [Thu, 11 May 2000 16:22:38 +0000 (16:22 +0000)]
Abuse .Ft and .Fn for macros in the usual way (name the macro parameters,
and don't use .It for the macro pseudo-return type).

24 years ago"Fixed" missing include in synopsis. POSIX.1-1996 only specifies
Bruce Evans [Thu, 11 May 2000 16:13:18 +0000 (16:13 +0000)]
"Fixed" missing include in synopsis.  POSIX.1-1996 only specifies
including <signal.h>, but that must be a bug in POSIX.1, because it
also specifies that the relevant prototype is [only] in <pthread.h>.

24 years agoFixed misspelling of a struct tag in a function parameter type.
Bruce Evans [Thu, 11 May 2000 16:03:38 +0000 (16:03 +0000)]
Fixed misspelling of a struct tag in a function parameter type.

24 years agoFixed missing include in synopsis.
Bruce Evans [Thu, 11 May 2000 16:01:17 +0000 (16:01 +0000)]
Fixed missing include in synopsis.

Use a long line instead splitting a line with backslash-newline in synopsis.
My synopsis checker doesn't understand backslash-newline.

24 years agoAdd workaround for ftpds with the Y2K MDTM bug
Dag-Erling Smørgrav [Thu, 11 May 2000 16:01:03 +0000 (16:01 +0000)]
Add workaround for ftpds with the Y2K MDTM bug

24 years agoFixed missing consts for function parameters, so that the code matches
Bruce Evans [Thu, 11 May 2000 15:57:17 +0000 (15:57 +0000)]
Fixed missing consts for function parameters, so that the code matches
the man page and POSIX.1.  Fixed nearby misformatting.  Fixed a missing
prototype.

24 years agoBump major number, since struct url has changed.
Dag-Erling Smørgrav [Thu, 11 May 2000 15:29:26 +0000 (15:29 +0000)]
Bump major number, since struct url has changed.

Discussed with: jdp

24 years agoFix hard sentence break.
Dag-Erling Smørgrav [Thu, 11 May 2000 15:22:52 +0000 (15:22 +0000)]
Fix hard sentence break.

Submitted by: sheldonh

24 years agoUnwind state on malloc() failure more carefully: avoids memory leak.
Tim Vanderhoek [Thu, 11 May 2000 14:57:26 +0000 (14:57 +0000)]
Unwind state on malloc() failure more carefully: avoids memory leak.

24 years agoSome white-box (NT) systems have SRM which reports a systype that's
Andrew Gallatin [Thu, 11 May 2000 13:31:05 +0000 (13:31 +0000)]
Some white-box (NT) systems have SRM which reports a systype that's
the negative of their blue-box (UNIX/OVMS) counterpart.  This was
causing us to panic early in the boot process because we weren't
expecting a negative index into the cpuinit[] array.

Obtained from: NetBSD
Reported by: Brett Bump <bbump@mail.enetis.net>

24 years agoReorganize some of the http code and split it into more functions.
Dag-Erling Smørgrav [Thu, 11 May 2000 13:31:02 +0000 (13:31 +0000)]
Reorganize some of the http code and split it into more functions.
Implement fetchStatHTTP().
Unbungle struct url, and add fetchFreeURL().
Document it.

24 years agoSupported PC-98.
Yoshihiro Takahashi [Thu, 11 May 2000 11:09:50 +0000 (11:09 +0000)]
Supported PC-98.

24 years agoAllow user to specify which logs to rotate
Dag-Erling Smørgrav [Thu, 11 May 2000 10:17:54 +0000 (10:17 +0000)]
Allow user to specify which logs to rotate

24 years agoAdd the '-l' option which prints string attribute followed by a file name.
Boris Popov [Thu, 11 May 2000 10:00:22 +0000 (10:00 +0000)]
Add the '-l' option which prints string attribute followed by a file name.

Reviewed by: rwatson, sheldonh

24 years agoAdd external for -i option.
Greg Lehey [Thu, 11 May 2000 08:57:09 +0000 (08:57 +0000)]
Add external for -i option.

24 years agoFix a typo
Brian Somers [Thu, 11 May 2000 07:59:21 +0000 (07:59 +0000)]
Fix a typo

24 years agoMention that the default is to let external traffic route to
Brian Somers [Thu, 11 May 2000 07:55:44 +0000 (07:55 +0000)]
Mention that the default is to let external traffic route to
the internal network when NAT is enabled.

Allow ``set target MYADDR'' to stop packets at the gateway.

24 years agoRevert the default behaviour for incoming connections so
Brian Somers [Thu, 11 May 2000 07:52:21 +0000 (07:52 +0000)]
Revert the default behaviour for incoming connections so
that they (once again) go to the target machine rather than
the alias address.

PR: 18354
Submitted by: ru

24 years agoRemove description of the -v option for init and start.
Greg Lehey [Thu, 11 May 2000 07:34:54 +0000 (07:34 +0000)]
Remove description of the -v option for init and start.

Describe -i option for start.

Update description of start command (will now work on plexes).

Update discussion of choice of stripe size (don't use power of 2).

Remove references to raw devices.

24 years agoAdd -i (interval) option, for use with start command.
Greg Lehey [Thu, 11 May 2000 07:33:53 +0000 (07:33 +0000)]
Add -i (interval) option, for use with start command.

24 years agovinum_info: Use the correct field to determine whether trace entries
Greg Lehey [Thu, 11 May 2000 07:32:09 +0000 (07:32 +0000)]
vinum_info: Use the correct field to determine whether trace entries
are read or write.

Bug-introduced-in: Revision 1.26

Show percentage complete in brief status of reviving and initializing
subdisks.

24 years agostart plex/subdisk:
Greg Lehey [Thu, 11 May 2000 07:29:09 +0000 (07:29 +0000)]
start plex/subdisk:
   Add -i option to delay between blocks copied, to reduce performance
   impact.

   Remove -v option, things work now.

Rewrite parity maintenance functions.

Rebuildparity: Add -v option to check whether it was broken before
rebuilding each block.

24 years agoAllow revive block sizes up to MAXPHYS.
Greg Lehey [Thu, 11 May 2000 07:27:27 +0000 (07:27 +0000)]
Allow revive block sizes up to MAXPHYS.

Change struct plex to reflect changes to parity maintenance functions.

24 years agoRewrite parity maintenance functions.
Greg Lehey [Thu, 11 May 2000 07:26:33 +0000 (07:26 +0000)]
Rewrite parity maintenance functions.

revive_block: Correct bug introduced in revision 1.25 which caused
Add fields to vinum_ioctl_msgexcessive concurrent requests followed by
system death.

24 years agoCosmetics.
Greg Lehey [Thu, 11 May 2000 07:25:36 +0000 (07:25 +0000)]
Cosmetics.

24 years agovinum_scandisk: Correct error recovery if a configuration is munged.
Greg Lehey [Thu, 11 May 2000 07:24:41 +0000 (07:24 +0000)]
vinum_scandisk: Correct error recovery if a configuration is munged.

24 years agoCorrect a comment.
Greg Lehey [Thu, 11 May 2000 07:23:50 +0000 (07:23 +0000)]
Correct a comment.

24 years agoChange parity maintenance ioctls.
Greg Lehey [Thu, 11 May 2000 07:23:16 +0000 (07:23 +0000)]
Change parity maintenance ioctls.

Add fields to vinum_ioctl_msg.

24 years agoChange linkage to parity maintenance operations.
Greg Lehey [Thu, 11 May 2000 07:22:18 +0000 (07:22 +0000)]
Change linkage to parity maintenance operations.

24 years agoPedantic change: "reboot" -> "boot"
David Nugent [Thu, 11 May 2000 06:31:59 +0000 (06:31 +0000)]
Pedantic change: "reboot" -> "boot"

24 years agoWhen "any" acts as a subject, the verb must agree with whatever any is of.
Tim Vanderhoek [Thu, 11 May 2000 05:29:10 +0000 (05:29 +0000)]
When "any" acts as a subject, the verb must agree with whatever any is of.

24 years agoContent-free commit: only remove trailing whitespace
Tim Vanderhoek [Thu, 11 May 2000 05:06:00 +0000 (05:06 +0000)]
Content-free commit: only remove trailing whitespace

24 years agoThe accept() function is a call, not an argument. Also, add: serial
Tim Vanderhoek [Thu, 11 May 2000 05:04:30 +0000 (05:04 +0000)]
The accept() function is a call, not an argument.  Also, add: serial
comma, missing-hyphen, and a word-erase character.

24 years agoMove <machine/ipl.h> outside #ifdef SMP because it supplies AST_RESCHED.
Peter Wemm [Thu, 11 May 2000 02:32:36 +0000 (02:32 +0000)]
Move <machine/ipl.h> outside #ifdef SMP because it supplies AST_RESCHED.
Without this, it shows up as an undefined symbol in /kernel. (!)
(This looks very freaky when doing a nm /kernel!)

24 years agoAttempt to work around problems caused by spurious interrupts and
Mike Smith [Thu, 11 May 2000 01:12:46 +0000 (01:12 +0000)]
Attempt to work around problems caused by spurious interrupts and
uninitialised interrupts in the APIC.  This seems to fix the problems
being seen on systems using the RCC chipsets, eg. Dell PowerEdge 24x0.

The actual nature of the problem probably needs further investigation,
but this patch allows us to actually function on these systems.

Submitted by: Drew Eckhardt <drew@Poohsticks.Org>

24 years agoMake the TAB key tab. Under syscons, shift-TAB also performs a backtab.
Tim Vanderhoek [Thu, 11 May 2000 00:56:55 +0000 (00:56 +0000)]
Make the TAB key tab.  Under syscons, shift-TAB also performs a backtab.

24 years agoFix a comment and actually expand \t the way the comment suggests.
Tim Vanderhoek [Thu, 11 May 2000 00:46:50 +0000 (00:46 +0000)]
Fix a comment and actually expand \t the way the comment suggests.