]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoFixed editing error in previous commit. `-lmd' (link libmd) became `-md'
bde [Sun, 15 Aug 1999 13:14:50 +0000 (13:14 +0000)]
Fixed editing error in previous commit.  `-lmd' (link libmd) became `-md'
(emulate the 'd' linker (?)).  This was most harmful for the NOSHARED=yes
case since libskey.o isn't linked to libmd.a.

Fixed the usual disorder of DPADD and LDADD, and some tab lossage.

24 years agoMinor mdoc fix.
mpp [Sun, 15 Aug 1999 12:16:59 +0000 (12:16 +0000)]
Minor mdoc fix.

24 years agoMinor style fix.
mpp [Sun, 15 Aug 1999 12:16:12 +0000 (12:16 +0000)]
Minor style fix.

24 years agoVarious man page cleanup:
mpp [Sun, 15 Aug 1999 10:48:36 +0000 (10:48 +0000)]
Various man page cleanup:

- Sort xrefs
- FreeBSD.ORG -> FreeBSD.org
- Be consistent with section names as outlined in mdoc(7).
- Other misc mdoc cleanup.

PR: doc/13144
Submitted by: Alexey M. Zelkin <phantom@cris.net>

24 years agoVarious man page cleanup:
mpp [Sun, 15 Aug 1999 10:01:15 +0000 (10:01 +0000)]
Various man page cleanup:

- Be consistent with section names as outlined in mdoc(7).
- Other misc mdoc cleanup.

24 years agoGive if_tun the "almost clone" makeover.
phk [Sun, 15 Aug 1999 09:54:57 +0000 (09:54 +0000)]
Give if_tun the "almost clone" makeover.

24 years agoVarious man page cleanup:
mpp [Sun, 15 Aug 1999 09:51:25 +0000 (09:51 +0000)]
Various man page cleanup:

- Sort xrefs
- FreeBSD.ORG -> FreeBSD.org
- Be consistent with section names as outlines in mdoc(7)
- Other misc mdoc cleanup.

PR: doc/13144
Submitted by: Alexy M. Zelkin <phantom@cris.net>

24 years agoGive BPF the "almost-clone" update. If you need more of them, make
phk [Sun, 15 Aug 1999 09:38:21 +0000 (09:38 +0000)]
Give BPF the "almost-clone" update.  If you need more of them, make
more entries in /dev and be happy you don't need to recompile your
kernel.

24 years agoIntroduce lminor(dev_t dev), which returns a linear minor number,
phk [Sun, 15 Aug 1999 09:32:47 +0000 (09:32 +0000)]
Introduce lminor(dev_t dev), which returns a linear minor number,
ie: hides the fact that the major number is stuck in the middle.

24 years agoVarious man page cleanup:
mpp [Sun, 15 Aug 1999 09:06:06 +0000 (09:06 +0000)]
Various man page cleanup:

- Sort xrefs
- Be consistent with section names as outlined in mdoc(7).
- Other misc mdoc cleanup.

PR: doc/13144
Submitted by: Alexey M. Zelkin <phantom@cris.net>

24 years agoVarious man page cleanup:
mpp [Sun, 15 Aug 1999 08:47:11 +0000 (08:47 +0000)]
Various man page cleanup:

- Sort xrefs.
- FreeBSD.ORG -> FreeBSD.org
- Be consistent with section names as outlined in mdoc(7).
- Other misc mdoc cleanup.

PR: doc/13144
Submitted by: Alexey M. Zelkin <phantom@cris.net>

24 years agoVarious man page cleanup:
mpp [Sun, 15 Aug 1999 08:25:34 +0000 (08:25 +0000)]
Various man page cleanup:

- Sort xrefs
- Be consistent with section names as outlines in mdoc(7).
- Other misc mdoc cleanup.

PR: doc/13144
Submitted by: Alexey M. Zelkin <phantom@cris.net>

24 years agoVarious man page cleanup:
mpp [Sun, 15 Aug 1999 08:06:12 +0000 (08:06 +0000)]
Various man page cleanup:

- Sort xrefs
- FreeBSD.ORG -> FreeBSD.org
- Properly xref RFCs.
- Be consistent with section names as outlined in mdoc(7).
- Other misc mdoc cleanup.

PR: doc/13144
Submitted by: Alexey M. Zelkin <phantom@cris.net>

24 years agoCorrectly save `flags' bits.
yokota [Sun, 15 Aug 1999 06:06:14 +0000 (06:06 +0000)]
Correctly save `flags' bits.

24 years agocontigmalloc1 (currently) depends on PQ_FREE and PQ_CACHE not being 0
alc [Sun, 15 Aug 1999 05:36:43 +0000 (05:36 +0000)]
contigmalloc1 (currently) depends on PQ_FREE and PQ_CACHE not being 0
to tell a valid "struct vm_page" from an invalid one in the vm_page_array.
This isn't a very robust method.

24 years agoinitsd: On completion, put the subdisk in 'initialized' status, not
grog [Sun, 15 Aug 1999 02:32:19 +0000 (02:32 +0000)]
initsd: On completion, put the subdisk in 'initialized' status, not
'up'.  This matches changes in the kernel module.

Reported-by: Bernd Walter <ticso@cicely.de>
Remy Nonnenmacher <remy@synx.com>

24 years agoAdd note that a change needs backing out.
grog [Sun, 15 Aug 1999 02:31:19 +0000 (02:31 +0000)]
Add note that a change needs backing out.

24 years agoEnsure that we can't create two different drives on the same
grog [Sun, 15 Aug 1999 02:30:34 +0000 (02:30 +0000)]
Ensure that we can't create two different drives on the same
partition.

Reported-by: Remy Nonnenmacher <remy@synx.com>

24 years agoAdd 'initialized' state for subdisks. After initializing, the subdisk
grog [Sun, 15 Aug 1999 02:29:14 +0000 (02:29 +0000)]
Add 'initialized' state for subdisks.  After initializing, the subdisk
goes into initialized state, not 'up'.  This makes it easier to ensure
consistency in multi-plex volumes.

update_plex_state: redo transitions from empty and initialized
subdisks to up or reviving, depending on the number of plexes.

Reported-by: Bernd Walter <ticso@cicely.de>
Remy Nonnenmacher <remy@synx.com>

24 years agoAdd back in old definitions if we're compiling for alpha.
mjacob [Sun, 15 Aug 1999 01:16:53 +0000 (01:16 +0000)]
Add back in old definitions if we're compiling for alpha.

24 years agoSight addendum to previous comment. The `undefined' token initialization
obrien [Sat, 14 Aug 1999 23:04:23 +0000 (23:04 +0000)]
Sight addendum to previous comment.  The `undefined' token initialization
bug was actually fixed in Bison 1.26.  My previous comment was in context
to our CVS repository.

24 years agoUse Bison with "--broken-undeftoken-init".
obrien [Sat, 14 Aug 1999 22:44:36 +0000 (22:44 +0000)]
Use Bison with "--broken-undeftoken-init".

24 years agoRather than deleting the grammer files during building, move them out of the
obrien [Sat, 14 Aug 1999 22:40:01 +0000 (22:40 +0000)]
Rather than deleting the grammer files during building, move them out of the
way.  This way they are available for inspection when debugging problems.

24 years agoDocument how I imported this.
obrien [Sat, 14 Aug 1999 22:33:55 +0000 (22:33 +0000)]
Document how I imported this.

24 years agoDue to brain damage on the FSF people, we can not longer pass "config.h"
obrien [Sat, 14 Aug 1999 22:32:36 +0000 (22:32 +0000)]
Due to brain damage on the FSF people, we can not longer pass "config.h"
type settings on the compiler command line.  FSF put their version of our
"__P()" macro in config.h, thus ``config.h'' required now.

24 years agoFixes for Bison 1.25->1.28 upgrade.
obrien [Sat, 14 Aug 1999 22:30:36 +0000 (22:30 +0000)]
Fixes for Bison 1.25->1.28 upgrade.

24 years agoBad reference vidcontrol(4) -> vidcontrol(1)
chris [Sat, 14 Aug 1999 22:29:34 +0000 (22:29 +0000)]
Bad reference vidcontrol(4) -> vidcontrol(1)

24 years agoBad reference getsockname(8) -> getsockname(2)
chris [Sat, 14 Aug 1999 22:25:59 +0000 (22:25 +0000)]
Bad reference getsockname(8) -> getsockname(2)

24 years agoAdd "--broken-undeftoken-init" option. This makes Bison 1.28 bug compatable
obrien [Sat, 14 Aug 1999 22:15:21 +0000 (22:15 +0000)]
Add "--broken-undeftoken-init" option.  This makes Bison 1.28 bug compatable
with version 1.25.

Bison 1.28 fixed a bug in the initialization of the `undefined' symbol
table entry.  Unfortunately something about the way we are compiling
egcs-1.1.2's cc1plus breaks with this bugfix.

"--broken-undeftoken-init" is needed when compiling cc1plus.  Otherwise
/usr/libexec/cc1plus cannot compile libgcc.

24 years agoFix conflicts.
obrien [Sat, 14 Aug 1999 21:59:04 +0000 (21:59 +0000)]
Fix conflicts.

24 years agoFix conflicts, ensure rev 1.2 changes are made to v1.28.
obrien [Sat, 14 Aug 1999 21:56:35 +0000 (21:56 +0000)]
Fix conflicts, ensure rev 1.2 changes are made to v1.28.

24 years agoUse stock v1.28 bison.simple.
obrien [Sat, 14 Aug 1999 21:53:50 +0000 (21:53 +0000)]
Use stock v1.28 bison.simple.
Bison 1.28 has our bison.simple rev 1.2 change, but due to code movement,
``cvs'' didn't realize this.

24 years agoThis commit was generated by cvs2svn to compensate for changes in r49795,
obrien [Sat, 14 Aug 1999 21:39:07 +0000 (21:39 +0000)]
This commit was generated by cvs2svn to compensate for changes in r49795,
which included commits to RCS files with non-trunk default branches.

24 years agoVirgin import of GNU Bison 1.28
obrien [Sat, 14 Aug 1999 21:39:07 +0000 (21:39 +0000)]
Virgin import of GNU Bison 1.28

24 years agoBad reference of uname(2) -> uname(3)
chris [Sat, 14 Aug 1999 21:37:28 +0000 (21:37 +0000)]
Bad reference of uname(2) -> uname(3)

24 years agoFix bad references, remove some invalid ones such as sa(9).
chris [Sat, 14 Aug 1999 20:51:13 +0000 (20:51 +0000)]
Fix bad references, remove some invalid ones such as sa(9).

24 years agoFix potential overflow, remove unnecessary bzero.
alfred [Sat, 14 Aug 1999 19:58:58 +0000 (19:58 +0000)]
Fix potential overflow, remove unnecessary bzero.
Pointed out by: green

remove redundant strlen, sprintf returns the length.

Reviewed by: peter

24 years agoChange to continue process after calling `Get Resume TImer' APM
iwasaki [Sat, 14 Aug 1999 18:45:19 +0000 (18:45 +0000)]
Change to continue process after calling `Get Resume TImer' APM
BIOS function (supported in v1.2) on older BIOS version (v1.1 or 1.0).

Reviewed by: -mobile ML folks.

24 years ago- Add apm_check_function_supported() and call it from
iwasaki [Sat, 14 Aug 1999 18:39:40 +0000 (18:39 +0000)]
- Add apm_check_function_supported() and call it from
  apm_bioscall() to check requested BIOS is supported or not.

- Add workaround in apm_driver_version() for the buggy BIOSes which
  don't return the connection version in %ax.

PR: i386/13028
Reviewed by: sanpei@sanpei.org and Warner Losh.

24 years agoImplementation of the linux_getcwd syscall.
marcel [Sat, 14 Aug 1999 18:20:44 +0000 (18:20 +0000)]
Implementation of the linux_getcwd syscall.

24 years agoExtend the number of syscalls to include those present in Linux 2.2.10. These
marcel [Sat, 14 Aug 1999 18:19:00 +0000 (18:19 +0000)]
Extend the number of syscalls to include those present in Linux 2.2.10. These
are syscalls 183 to 190. Also implement syscall 183: linux_getcwd. This is
needed to support a RH 6.0 environment.

24 years agoImplementation of linux_rt_sigaction and linux_rt_sigprocmask syscalls. Both
marcel [Sat, 14 Aug 1999 17:28:46 +0000 (17:28 +0000)]
Implementation of linux_rt_sigaction and linux_rt_sigprocmask syscalls. Both
functions use the new sigset_t and sigaction_t which allows support for more
than 32 signals. Only the lower 32 signals are supported for now.

linux_rt_sigaction, linux_sigaction and linux_signal use linux_do_sigaction
to do the actual work. That way unnecessary redundancy is avoided. The same
has been done for linux_rt_sigprocmask and linux_sigprocmask. They call
linux_do_sigprocmask to do the actual work.

24 years agos/Yes/YES/g. `tcp_extensions''s comment implied "Yes" was a valid value
obrien [Sat, 14 Aug 1999 17:15:19 +0000 (17:15 +0000)]
s/Yes/YES/g.  `tcp_extensions''s comment implied "Yes" was a valid value
for turning on the feature.  Which is wrong.

Submitted by: bde

24 years agoImplementation of linux_rt_sigaction and linux_rt_sigprocmask syscalls. These
marcel [Sat, 14 Aug 1999 17:14:13 +0000 (17:14 +0000)]
Implementation of linux_rt_sigaction and linux_rt_sigprocmask syscalls. These
syscalls are needed to support a RH 6.0 environment.

24 years agoHopefully clarify "log_in_vain" knob's comment.
obrien [Sat, 14 Aug 1999 17:13:40 +0000 (17:13 +0000)]
Hopefully clarify "log_in_vain" knob's comment.

Reviewed by: bde, sheldonh, andreas

24 years agoAppend the flags from the "boot" command to those which came from
jdp [Sat, 14 Aug 1999 17:06:11 +0000 (17:06 +0000)]
Append the flags from the "boot" command to those which came from
the SRM environment.  This makes the traditional "boot [/kernel] -s"
and similar things work on the Alpha.  Since the flags are appended,
they augment and/or override those from the SRM environment.

24 years agoMinor glitch in ti_newbuf_jumbo(): m_adj() was being called on
wpaul [Sat, 14 Aug 1999 15:45:03 +0000 (15:45 +0000)]
Minor glitch in ti_newbuf_jumbo(): m_adj() was being called on
m instead of m_new.

Submitted by: Kenneth D. Merry <ken@kdm.org>

24 years agoDon't force compression in SUPFLAGS, since that effectively negates
sheldonh [Sat, 14 Aug 1999 14:05:40 +0000 (14:05 +0000)]
Don't force compression in SUPFLAGS, since that effectively negates
the functionality of the compression option in the supfile.

Reported by: Ben Rosengart <ben@skunk.org>

24 years agoadd debugging message in unregister_method
n_hibma [Sat, 14 Aug 1999 13:32:25 +0000 (13:32 +0000)]
add debugging message in unregister_method

24 years agoFix LINUX_TIOC{S|G}SERIAL implementation. Both do not copy data in or out
marcel [Sat, 14 Aug 1999 13:26:44 +0000 (13:26 +0000)]
Fix LINUX_TIOC{S|G}SERIAL implementation. Both do not copy data in or out
of kernel space. Remove the ioctl supporting functions, and move the actual
code to the switch-statement. Now everybody can clearly see that the
implementation is really poor.

Also fix a typo in LINUX_TIOCGETD. The underlying function was given command
TIOCSETD instead op TIOCGETD...

24 years agoMinor spelling glitch
n_hibma [Sat, 14 Aug 1999 13:20:04 +0000 (13:20 +0000)]
Minor spelling glitch

24 years agoReviewed by: phk
pho [Sat, 14 Aug 1999 12:19:40 +0000 (12:19 +0000)]
Reviewed by: phk
When fts_open is used with option FTS_NOCHDIR the full
path entry of type FTS_DP is returned with a trailing
'/' if the final directory is empty.
This fix coresponds to netbsd's __fts13.c v. 1.16

24 years agoSpring cleaning around strategy and disklabels/slices:
phk [Sat, 14 Aug 1999 11:40:51 +0000 (11:40 +0000)]
Spring cleaning around strategy and disklabels/slices:

Introduce BUF_STRATEGY(struct buf *, int flag) macro, and use it throughout.
please see comment in sys/conf.h about the flag argument.

Remove strategy argument from all the diskslice/label/bad144
implementations, it should be found from the dev_t.

Remove bogus and unused strategy1 routines.

Remove open/close arguments from dssize().  Pick them up from dev_t.

Remove unused and unfinished setgeom support from diskslice/label/bad144 code.

24 years agoAvoid possible panic by checking for EFAULT from copyinstr() during
newton [Sat, 14 Aug 1999 10:52:33 +0000 (10:52 +0000)]
Avoid possible panic by checking for EFAULT from copyinstr() during
pathname translation.

Submitted by: green

24 years agoFix the LINUX_TCSET{A|AW|AF} and LINUX_TCSET{S|SW|SF} ioctls. These all suffer
marcel [Sat, 14 Aug 1999 10:30:38 +0000 (10:30 +0000)]
Fix the LINUX_TCSET{A|AW|AF} and LINUX_TCSET{S|SW|SF} ioctls. These all suffer
from the same bug in that the argument is not first copied from user space
before it is used. This is part 2 (of 2) of the termios fixes.

24 years agoFix a couple of termio/termios conversion bugs/typos/dupos/brainos and other
marcel [Sat, 14 Aug 1999 10:05:50 +0000 (10:05 +0000)]
Fix a couple of termio/termios conversion bugs/typos/dupos/brainos and other
changes. This is part 1 of the complete termios ioctl fixes.

o  change type of c_{i|o|c|l}flag in struct termios from unsigned long to
   unsigned int. The type now matches the Linux definitions.
o  replaced constants by the corresponding defines in sptab[] for clarity.
   Since there's no define for 135 baud, its mapping has been dropped.

function bsd_to_linux_termios:
o  Fix typo IXON -> IXANY.
o  Remove bogus assignment to c_cc[LINUX_VSWTC].

function linux_to_bsd_termios:
o  Fix dupo LINUX_IXON -> LINUX_IXANY.
o  Add LINUX_CREAD mapping.
o  Fix typo IEXTEN -> LINUX_IEXTEN.

function linux_to_bsd_termio:
o  Small optimization: Don't preset the complete c_cc array when we next
   assign to the first LINUX_NCC entries.

24 years agoDocument the AUTHORS section in mdoc(7). Expand the description
mpp [Sat, 14 Aug 1999 09:45:31 +0000 (09:45 +0000)]
Document the AUTHORS section in mdoc(7).  Expand the description
of the AUTHORS section in mdoc.samples(7) to document how the
authors name should be specified.

PR: docs/13131
Pointed out by: Alexey M. Zelkin <phantom@cris.net>

24 years agoCorrect typo.
grog [Sat, 14 Aug 1999 09:15:43 +0000 (09:15 +0000)]
Correct typo.

24 years agoBad reference to exit(2) changed to exit(3).
chris [Sat, 14 Aug 1999 08:13:35 +0000 (08:13 +0000)]
Bad reference to exit(2) changed to exit(3).

24 years agoBad reference exit(2) changed to exit(3)
chris [Sat, 14 Aug 1999 08:05:46 +0000 (08:05 +0000)]
Bad reference exit(2) changed to exit(3)

24 years agoAdd $Id$ tag.
chris [Sat, 14 Aug 1999 07:59:58 +0000 (07:59 +0000)]
Add $Id$ tag.

24 years agoBad reference time(2) changed to time(3)
chris [Sat, 14 Aug 1999 07:57:52 +0000 (07:57 +0000)]
Bad reference time(2) changed to time(3)

24 years agoBad reference of termios(3) changed to termios(4).
chris [Sat, 14 Aug 1999 07:52:29 +0000 (07:52 +0000)]
Bad reference of termios(3) changed to termios(4).

24 years agoBad reference of sysctl(1) changed to sysctl(8)
chris [Sat, 14 Aug 1999 07:46:50 +0000 (07:46 +0000)]
Bad reference of sysctl(1) changed to sysctl(8)

24 years agoBad reference to lstat(3) changed to lstat(2)
chris [Sat, 14 Aug 1999 07:43:02 +0000 (07:43 +0000)]
Bad reference to lstat(3) changed to lstat(2)

24 years agoFix .Xr line for `getpagesize'
chris [Sat, 14 Aug 1999 07:33:15 +0000 (07:33 +0000)]
Fix .Xr line for `getpagesize'

24 years agoFix some bad references:
chris [Sat, 14 Aug 1999 07:07:46 +0000 (07:07 +0000)]
Fix some bad references:
  fopen(2) -> fopen(3)
  fclose(2) -> fclose(3)

24 years agoFix bad references:
chris [Sat, 14 Aug 1999 06:52:28 +0000 (06:52 +0000)]
Fix bad references:
  vmstat(1) -> vmstat(8)
  iostat(1) -> iostat(8)

24 years agoBad reference of getcwd(2) changed to getcwd(3).
chris [Sat, 14 Aug 1999 06:47:13 +0000 (06:47 +0000)]
Bad reference of getcwd(2) changed to getcwd(3).

24 years agoBad reference of setrlimit(3) changed to setrlimit(2).
chris [Sat, 14 Aug 1999 06:39:48 +0000 (06:39 +0000)]
Bad reference of setrlimit(3) changed to setrlimit(2).

24 years agoBad reference of su(8) changed to su(1).
chris [Sat, 14 Aug 1999 06:34:50 +0000 (06:34 +0000)]
Bad reference of su(8) changed to su(1).

24 years agoTidy up comments.
grog [Sat, 14 Aug 1999 06:31:37 +0000 (06:31 +0000)]
Tidy up comments.

Add definitions for VINUM_BLOCK_PLEX and VINUM_CHAR_PLEX.

struct plex: add lock-related variables.

struct rangelock: add variables.

24 years agoClean up comments.
grog [Sat, 14 Aug 1999 06:30:56 +0000 (06:30 +0000)]
Clean up comments.

Don't return "can't do it" when the user requests a state change to
the current state.  This previously caused silly messages like "Can't
start <foo>: invalid argument", when in fact <foo> was already
started.

set_plex_state: don't set state for non-existent plexes.

update_plex_status: as long as we have initializing subdisks, we're
    initializing.

24 years agoClean up some comments.
grog [Sat, 14 Aug 1999 06:30:15 +0000 (06:30 +0000)]
Clean up some comments.

Move the declaration of freerq() to request.h.

logrq: add support for lock events.

vinumstart: solve a problem where removing a plex from an active
    volume could cause attempts to access non-existent plexes.

launch_requests: don't set a request group active until we're sure we
 can launch it.  This caused some hangs under unusual
 circumstances.

bre: don't set XFR_BAD_SUBDISK if we're not going to use it.

build_read_request: correct recovery, which caused some hangs under
    (other) unusual circumstances.

build_rq_buffer: don't set bp->b_dev if we don't have a dev.

sdio: clean up, remove obsolete code.

deallocrqg: unlock any locks the rqg may have.

24 years agoTidy up some comments.
grog [Sat, 14 Aug 1999 06:29:25 +0000 (06:29 +0000)]
Tidy up some comments.

bre5:
  Shorten some lines.

Desired-by: bde
  If we're reading from a short plex, return EOF indication.

  Always lock the stripe before starting a transfer.  Hopefully the
  current version will solve some data integrity problems that have
  been reported with degraded RAID-5 plexes.

  Reported-by: Bernd Walter <ticso@cicely.de>
Remy Nonnenmacher <remy@synx.com>

24 years agoAdd keypairs replace and readpol. These got forgotten in a previous
grog [Sat, 14 Aug 1999 06:28:48 +0000 (06:28 +0000)]
Add keypairs replace and readpol.  These got forgotten in a previous
commit, but since they don't work yet, nobody noticed.

24 years agorewrite lockrange and unlockrange. Hopefully the current version will
grog [Sat, 14 Aug 1999 06:28:21 +0000 (06:28 +0000)]
rewrite lockrange and unlockrange.  Hopefully the current version will
solve some data integrity problems that have been reported with
degraded RAID-5 plexes.

Reported-by: Bernd Walter <ticso@cicely.de>
Remy Nonnenmacher <remy@synx.com>

24 years agoAdd hotspare keyword.
grog [Sat, 14 Aug 1999 06:27:52 +0000 (06:27 +0000)]
Add hotspare keyword.

24 years agovinumioctl: add ioctls for raw subdisks and plexes. Remove code for
grog [Sat, 14 Aug 1999 06:27:25 +0000 (06:27 +0000)]
vinumioctl: add ioctls for raw subdisks and plexes.  Remove code for
drives (we refuse to touch drives, that's the disk driver's business).

24 years agoRemove some XXX comments (they're not really gone, just invisible).
grog [Sat, 14 Aug 1999 06:26:32 +0000 (06:26 +0000)]
Remove some XXX comments (they're not really gone, just invisible).
Tidy other comments.

open_drive: don't call set_drive_state if we decide to take it down.
This could help avoid some race conditions with the daemon.

init_drive: don't set the drive down, we'll let close_locked_drive do
that.

close_locked_drive: set drive state to down without calling
set_drive_state.  This could help avoid some race conditions with the
daemon.

driveio: remove the function, it wasn't being used.

get_volume_label: remove volume dependencies so that we can return a
label for plexes and subdisks as well.  What a kludge.

24 years agoBad reference of mount(1) changed to mount(8).
chris [Sat, 14 Aug 1999 06:26:14 +0000 (06:26 +0000)]
Bad reference of mount(1) changed to mount(8).

24 years agoDon't create a "struct vpgqueues" for PQ_NONE.
alc [Sat, 14 Aug 1999 06:25:54 +0000 (06:25 +0000)]
Don't create a "struct vpgqueues" for PQ_NONE.

24 years agoRemove some XXX comments (they're not really gone, just invisible).
grog [Sat, 14 Aug 1999 06:25:52 +0000 (06:25 +0000)]
Remove some XXX comments (they're not really gone, just invisible).

Remove declarations for freerq and free_rqg.

Remove DEBUG_RESID code.

freerq: check whether the request is holding a lock, free if so.

free_rqg: remove.  It wasn't being used any more.

Change the Debugger calls to panics.

24 years agoChange definitions for get_volume_label, lockrange.
grog [Sat, 14 Aug 1999 06:25:14 +0000 (06:25 +0000)]
Change definitions for get_volume_label, lockrange.

Move definition for unlockrange to request.h (it now contains
request-related parameters).

Add sddownstate.

24 years agoAdd code which someday will help support hot spares.
grog [Sat, 14 Aug 1999 06:24:34 +0000 (06:24 +0000)]
Add code which someday will help support hot spares.

checkdiskconfig(): remove.  It didn't make any sense to complain about
kernel keywords in user config files; it just made it more difficult
to convert.  Now we ignore kernel keywords if we're not in kernel
mode.

get_empty_sd: initialize sectors.

free_drive: don't close if we don't have a vp.  Maybe this will help
fix the problem that peter had, but I wouldn't count on it.

config_plex: If the plex is RAID-5, give it a rangelock structure.

start_config: Reset current drive, plex and volume so that a new
'create' command doesn't get long-dead defaults.

24 years agovinumopen: add code for opening raw plexes. Volume-related plex
grog [Sat, 14 Aug 1999 06:22:15 +0000 (06:22 +0000)]
vinumopen: add code for opening raw plexes.  Volume-related plex
devices will probably die a death soon (yes, DES, that's correct
usage).

24 years agoClarify some comments.
grog [Sat, 14 Aug 1999 06:21:45 +0000 (06:21 +0000)]
Clarify some comments.

struct rqelement, enum rqinfo_type, struct rqinfo, union rqinfou: add
lock requests.

Add declarations for freerq and unlockrange.  Since they include
request structures, they can't go in vinumext.h

24 years agoAdd declarations for initvol, initplex and initsd helper functions for
grog [Sat, 14 Aug 1999 06:20:55 +0000 (06:20 +0000)]
Add declarations for initvol, initplex and initsd helper functions for
vinum_init.

24 years agovinum_info: add code to print out lock events.
grog [Sat, 14 Aug 1999 06:20:20 +0000 (06:20 +0000)]
vinum_info: add code to print out lock events.
    modify format of other event printouts.

24 years agomake_plex_dev: make them raw. Volume-related plex devices will
grog [Sat, 14 Aug 1999 06:19:38 +0000 (06:19 +0000)]
make_plex_dev: make them raw.  Volume-related plex devices will
probably die a death soon (yes, DES, that's correct usage).

24 years agovinum_init: add code to initialize subdisks as well as plexes. Also
grog [Sat, 14 Aug 1999 06:18:39 +0000 (06:18 +0000)]
vinum_init: add code to initialize subdisks as well as plexes.  Also
provide for initializing volumes; this code is not yet complete.

24 years agoChange reference to mount(1) to mount(8)
chris [Sat, 14 Aug 1999 06:17:24 +0000 (06:17 +0000)]
Change reference to mount(1) to mount(8)

24 years agoMinor style fix - change 'if(!*v)' to 'if (!*v)'
chris [Sat, 14 Aug 1999 05:38:04 +0000 (05:38 +0000)]
Minor style fix - change 'if(!*v)' to 'if (!*v)'

24 years agoGrrrr. Fix a really lame bug that I tripped over while testing my miibus
wpaul [Sat, 14 Aug 1999 05:11:01 +0000 (05:11 +0000)]
Grrrr. Fix a really lame bug that I tripped over while testing my miibus
stuff: unregister_methods() is horribly broken. The idea, if I'm not mistaken,
is that the refcount on a method is decremented, and only when it reaches
zero is the method freed. However desc->method is set to NULL unconditionally
regardless of the refcount, which means the method pointer is trashed the
first time the method is deallocated. The obvious detrimental effect is
that memory is leaked. The not so obvious effect is that when you call
unregister_method() the second time on the same method, you get a NULL
pointer dereference and a panic.

Now I can successfully unload network device drivers and the miibus module
without crashing the system.

*sigh*

24 years agoBring up to date to work with -CURRENT.
grog [Sat, 14 Aug 1999 04:56:14 +0000 (04:56 +0000)]
Bring up to date to work with -CURRENT.

24 years agoUse proper capitalization.
obrien [Sat, 14 Aug 1999 03:42:27 +0000 (03:42 +0000)]
Use proper capitalization.

24 years agoUse better comment for tcp_keepalive option.
obrien [Sat, 14 Aug 1999 03:41:56 +0000 (03:41 +0000)]
Use better comment for tcp_keepalive option.

Submitted by: imp

24 years agoSpecify that `ruptime' data is broadcast once every three minutes, not
chris [Sat, 14 Aug 1999 01:04:10 +0000 (01:04 +0000)]
Specify that `ruptime' data is broadcast once every three minutes, not
once every minute.

PR: docs/13130
Submitted by: Ben Smithurst <ben@scientia.demon.co.uk>

24 years agoChange reference from kldload(3) to kldload(2)
chris [Fri, 13 Aug 1999 21:02:39 +0000 (21:02 +0000)]
Change reference from kldload(3) to kldload(2)

24 years agoAdd a notice saying that FreeBSD no longer tracks the PPS-API WG's drafts.
phk [Fri, 13 Aug 1999 20:56:35 +0000 (20:56 +0000)]
Add a notice saying that FreeBSD no longer tracks the PPS-API WG's drafts.