]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
25 years agoProperly install the *.h files in CORE and Config.pm where it wants to be.
Mark Murray [Thu, 15 Oct 1998 18:50:04 +0000 (18:50 +0000)]
Properly install the *.h files in CORE and Config.pm where it wants to be.

This should fix a substantial number of broken p5-* ports.

Tested-on: x11-toolkits/p5-Tk (Amongst others)

25 years agoBring back support for honoring the primary channel setting for twin channel
Justin T. Gibbs [Thu, 15 Oct 1998 18:21:50 +0000 (18:21 +0000)]
Bring back support for honoring the primary channel setting for twin channel
EISA adapters.  This could have caused the system to find the wrong root
disk.

Reviewed by:  J Wunsch <j@ida.interface-business.de>

25 years agoClean up some unused variables.
Kenneth D. Merry [Thu, 15 Oct 1998 17:46:26 +0000 (17:46 +0000)]
Clean up some unused variables.

Reviewed by: ken
Submitted by: phk

25 years ago- bzero() after malloc() bug fix
Peter Wemm [Thu, 15 Oct 1998 17:16:24 +0000 (17:16 +0000)]
- bzero() after malloc() bug fix
- look up relocation symbol names in correct table bugfix.
- remove unused (initialized to 0) variable and conditional free() of it.

25 years agoFix style issues in execl(), and make execle() vfork()-safe.
Dag-Erling Smørgrav [Thu, 15 Oct 1998 17:14:15 +0000 (17:14 +0000)]
Fix style issues in execl(), and make execle() vfork()-safe.

Reviewed by: bde
Approved by: jkh

25 years ago- bzero() after malloc(). This is especially obvious when kern_malloc is
Peter Wemm [Thu, 15 Oct 1998 17:12:12 +0000 (17:12 +0000)]
- bzero() after malloc().  This is especially obvious when kern_malloc is
compiled with DIAGNOSTIC.
- Don't break from the preload module processing loop prematurely.

25 years agoFix sysinit_add().
Peter Wemm [Thu, 15 Oct 1998 17:09:19 +0000 (17:09 +0000)]
Fix sysinit_add().
- Don't include multiple copies of the previous sysinit in the new one.
- Leave space for and explicitly null terminate the new list.

25 years agoGet the last used address via a more conservative method, don't depend
Peter Wemm [Thu, 15 Oct 1998 17:06:36 +0000 (17:06 +0000)]
Get the last used address via a more conservative method, don't depend
on the module chain being in increasing address order.

25 years agoFix make world breakage from last commit.
John Polstra [Thu, 15 Oct 1998 16:09:56 +0000 (16:09 +0000)]
Fix make world breakage from last commit.

25 years agoFixed breakage of gdbreplay's and gdbserver's BINDIR in previous commit.
Bruce Evans [Thu, 15 Oct 1998 14:15:09 +0000 (14:15 +0000)]
Fixed breakage of gdbreplay's and gdbserver's BINDIR in previous commit.
They may belong in /usr/libexec/{aout,elf}, but objformat doesn't
support that.

Fixed bogus `?=' assignments for BINDIR.

25 years agoSpelling.
Stephen McKay [Thu, 15 Oct 1998 13:37:35 +0000 (13:37 +0000)]
Spelling.

25 years agoatq ignored locale info when printing dates. Is the other call to
Stephen McKay [Thu, 15 Oct 1998 13:30:48 +0000 (13:30 +0000)]
atq ignored locale info when printing dates.  Is the other call to
setlocale() just in the wrong place?

PR: 8300

25 years agoFix a serious bug in make(1)'s handling of archive libraries.
Joseph Koshy [Thu, 15 Oct 1998 13:00:34 +0000 (13:00 +0000)]
Fix a serious bug in make(1)'s handling of archive libraries.

This case did not need to be tested when RANLIBMAG was defined
(as when in an `aout' environment) because Arch_StatMember() treated
the two cases of the library not being present and a member of
the library not being present the same way, forcing a rebuild
of the library.  Since in the ELF environment we don't look inside
archive libraries we now need to check if the archive library is
present in order to determine its `out-of-date'-ness.

(I hope I've been able to meet the Oct 15th freeze).

Reported-by: Steve Price (and a few others whom I've forgotten, sorry)

25 years agoMake it clear that installations can only occur from drive C.
Jordan K. Hubbard [Thu, 15 Oct 1998 11:27:45 +0000 (11:27 +0000)]
Make it clear that installations can only occur from drive C.
PR: 8334
Submitted by: Sue Blake <sue@vedanix.welearn.com.au>

25 years agoUpdate DOS emulator description.
Jordan K. Hubbard [Thu, 15 Oct 1998 11:24:29 +0000 (11:24 +0000)]
Update DOS emulator description.

25 years agoMake the bit about DOS partitions somewhat clearer.
Jordan K. Hubbard [Thu, 15 Oct 1998 11:22:33 +0000 (11:22 +0000)]
Make the bit about DOS partitions somewhat clearer.
PR: 8333

25 years agoMore attempts to DTRT for the alpha, based on what I now know about boot2.
Jordan K. Hubbard [Thu, 15 Oct 1998 11:12:40 +0000 (11:12 +0000)]
More attempts to DTRT for the alpha, based on what I now know about boot2.

25 years agoFix upgrade notes.
Jordan K. Hubbard [Thu, 15 Oct 1998 11:01:30 +0000 (11:01 +0000)]
Fix upgrade notes.
PR: 8331
Submitted by: Sue Blake <sue@vedanix.welearn.com.au>

25 years agoUpdate for 3.0
Jordan K. Hubbard [Thu, 15 Oct 1998 10:16:33 +0000 (10:16 +0000)]
Update for 3.0
Submitted by: Sue Blake <sue@vedanix.welearn.com.au>

25 years agoCall ourselves by the correct name. :)
Jordan K. Hubbard [Thu, 15 Oct 1998 10:14:55 +0000 (10:14 +0000)]
Call ourselves by the correct name. :)
Submitted by: Sue Blake <sue@vedanix.welearn.com.au>

25 years agoAdd ipfw change information.
Jordan K. Hubbard [Thu, 15 Oct 1998 10:13:53 +0000 (10:13 +0000)]
Add ipfw change information.

25 years agoUpdate size information for ports.
Jordan K. Hubbard [Thu, 15 Oct 1998 10:13:34 +0000 (10:13 +0000)]
Update size information for ports.

25 years agoSearch for correct property name; the older attrs code bogusly folded case,
Jordan K. Hubbard [Thu, 15 Oct 1998 10:03:48 +0000 (10:03 +0000)]
Search for correct property name; the older attrs code bogusly folded case,
so I never noticed this bug. :)

25 years agoChange a bogus cast to the correct one.
Doug Rabson [Thu, 15 Oct 1998 09:53:27 +0000 (09:53 +0000)]
Change a bogus cast to the correct one.

25 years agoDon't frob the user stack directly, use suword instead. This fixes the
Doug Rabson [Thu, 15 Oct 1998 09:52:19 +0000 (09:52 +0000)]
Don't frob the user stack directly, use suword instead.  This fixes the
elf_freebsd_fixup() panic which many people have noticed on the alpha.

25 years agoUse the interface name as well as the method name when creating the method
Doug Rabson [Thu, 15 Oct 1998 09:50:45 +0000 (09:50 +0000)]
Use the interface name as well as the method name when creating the method
description.

Submitted by: Nick Hibma <nick.hibma@jrc.it>

25 years agoMake the next_writeable address specific for each drive, there might
Søren Schmidt [Thu, 15 Oct 1998 08:11:55 +0000 (08:11 +0000)]
Make the next_writeable address specific for each drive, there might
be more than one burner online.

25 years agoUpdate README.TXT, not that anyone will read it anyway, to describe just
Jordan K. Hubbard [Thu, 15 Oct 1998 07:42:25 +0000 (07:42 +0000)]
Update README.TXT, not that anyone will read it anyway, to describe just
what 3.0 is meant to be.

25 years agosnuff warning message if package is not found in INDEX but is found
Jordan K. Hubbard [Thu, 15 Oct 1998 06:50:43 +0000 (06:50 +0000)]
snuff warning message if package is not found in INDEX but is found
to be already installed anyway.

25 years agoFix formatting bug.
Joseph Koshy [Thu, 15 Oct 1998 05:04:28 +0000 (05:04 +0000)]
Fix formatting bug.

PR: 8320
Submitted by: Amakawa Shuhei <amakawa@sf.t.u-tokyo.ac.jp>

25 years agoAdd lstat() as a known system call.
Sean Eric Fagan [Thu, 15 Oct 1998 04:31:44 +0000 (04:31 +0000)]
Add lstat() as a known system call.

25 years agoForgot to commit this; the alpha uses the kld symbol interface now. The
Peter Wemm [Thu, 15 Oct 1998 02:01:50 +0000 (02:01 +0000)]
Forgot to commit this; the alpha uses the kld symbol interface now.  The
tables that db_elf.c is expecting are not loaded in that format any more.

25 years agoConform to POSIX and close any copies of popen() descriptors inherited by a
Mike Smith [Thu, 15 Oct 1998 01:47:40 +0000 (01:47 +0000)]
Conform to POSIX and close any copies of popen() descriptors inherited by a
popen()ed child.

PR: misc/7810
Submitted by: Wayne Scott <wscott@ichips.intel.com>

25 years agoAdd objformat links for gdb and (while we're at it) c++filt.
Peter Wemm [Thu, 15 Oct 1998 00:19:08 +0000 (00:19 +0000)]
Add objformat links for gdb and (while we're at it) c++filt.
This means we can (assuming an a.out gdb is present) debug a.out kernels
and crashdumps with gdb -aout.

25 years agoInstall gdb in /usr/lib/aout or /usr/lib/elf depending on format.
Peter Wemm [Thu, 15 Oct 1998 00:15:01 +0000 (00:15 +0000)]
Install gdb in /usr/lib/aout or /usr/lib/elf depending on format.

25 years agoFix a couple of potential buffer overrun cases.
Kenneth D. Merry [Wed, 14 Oct 1998 23:28:26 +0000 (23:28 +0000)]
Fix a couple of potential buffer overrun cases.

Submitted by: imp

25 years agoNarrow the quirk entry for the Seagate Elite 9 a bit to just cover drives
Kenneth D. Merry [Wed, 14 Oct 1998 22:51:51 +0000 (22:51 +0000)]
Narrow the quirk entry for the Seagate Elite 9 a bit to just cover drives
with 71* firmware revisions.  Scott Mace <smace@intt.ORG> reports that
drives with 00* firmware revisions do tagged queueing just fine.

25 years ago- Add a couple comment lines to note that spaces are not allowed as
Nate Williams [Wed, 14 Oct 1998 21:59:55 +0000 (21:59 +0000)]
- Add a couple comment lines to note that spaces are not allowed as
  field separators.

PR: conf/8162
Submitted by: Sheldon Hearn <sheldonh@axl.training.iafrica.com>

25 years agoDisable tagged queueing for the Seagate Elite 9GB drives. They tend to get
Kenneth D. Merry [Wed, 14 Oct 1998 21:17:39 +0000 (21:17 +0000)]
Disable tagged queueing for the Seagate Elite 9GB drives.  They tend to get
hung up when you send tags to them too quickly.  (CAM is able to recover
from the problem, but this just avoids it altogether.)

Reviewed by: gibbs
Reported by: Bret Ford <bford@uop.cs.uop.edu>
and: Martin Renters <martin@tdc.on.ca>

25 years agoDisable the 'devstat_end_transaction' busy count printf until after 3.0
Kenneth D. Merry [Wed, 14 Oct 1998 20:44:05 +0000 (20:44 +0000)]
Disable the 'devstat_end_transaction' busy count printf until after 3.0
release goes out the door.  We know there's a bug in the devstat
implementation in the wd driver, but bde and msmith haven't been able to
fix it yet.

So, disable the printf to avoid confusing/worrying people.

Suggested by: msmith

25 years agoActivate boot2.
Robert Nordier [Wed, 14 Oct 1998 20:40:56 +0000 (20:40 +0000)]
Activate boot2.

25 years agoCorrect braino in previous commit. I get the pointy hat again.
Dag-Erling Smørgrav [Wed, 14 Oct 1998 20:23:40 +0000 (20:23 +0000)]
Correct braino in previous commit. I get the pointy hat again.

25 years agoMake execl() vfork()-safe. This should fix potential bugs in rcp,
Dag-Erling Smørgrav [Wed, 14 Oct 1998 18:53:36 +0000 (18:53 +0000)]
Make execl() vfork()-safe. This should fix potential bugs in rcp,
telnet and tip, and probably a few other apps.

Reviewed by: bde
Approved by: jkh

25 years agoRevert 1.33; lstat() is indeed the wrong thing to do here. I'll
Jordan K. Hubbard [Wed, 14 Oct 1998 18:52:04 +0000 (18:52 +0000)]
Revert 1.33; lstat() is indeed the wrong thing to do here.  I'll
revisit this after release.

25 years agoAndrew Gallatin reported some internal queue overflows with MAXISPREQUEST
Matt Jacob [Wed, 14 Oct 1998 16:55:00 +0000 (16:55 +0000)]
Andrew Gallatin reported some internal queue overflows with MAXISPREQUEST
at 256. So- to be safe, let's roll back to 64 while we do some more
investigating.

25 years agoCheck the executable's header to make sure it is a valid executable.
John Polstra [Wed, 14 Oct 1998 16:16:50 +0000 (16:16 +0000)]
Check the executable's header to make sure it is a valid executable.
If it is ELF, print a diagnostic saying that it is not supported yet
by this program.  This is a stop-gap anti-bug-report measure because
it looks like there won't be time to implement gcore's ELF support
before 3.0 is released.

25 years agoBacked out rev. 1.164. It caused problems on SMP.
Dmitrij Tejblum [Wed, 14 Oct 1998 15:05:52 +0000 (15:05 +0000)]
Backed out rev. 1.164. It caused problems on SMP.

PR: 8309

25 years agoLink fdboot to boot1 as well.
Peter Wemm [Wed, 14 Oct 1998 13:19:52 +0000 (13:19 +0000)]
Link fdboot to boot1 as well.

25 years agoUse properties code.
Jordan K. Hubbard [Wed, 14 Oct 1998 11:23:48 +0000 (11:23 +0000)]
Use properties code.

25 years agocorrect prototype.
Jordan K. Hubbard [Wed, 14 Oct 1998 11:04:36 +0000 (11:04 +0000)]
correct prototype.

25 years agoUse more rational inode sizes.
Jordan K. Hubbard [Wed, 14 Oct 1998 10:58:12 +0000 (10:58 +0000)]
Use more rational inode sizes.

25 years agoFixed 7 style bugs. The contents of such a simple Makefile is entirely
Bruce Evans [Wed, 14 Oct 1998 10:21:10 +0000 (10:21 +0000)]
Fixed 7 style bugs.  The contents of such a simple Makefile is entirely
determined by style rules, the program name and the existence of the man
page.

25 years agoFix a warning I missed before.
Peter Wemm [Wed, 14 Oct 1998 10:08:35 +0000 (10:08 +0000)]
Fix a warning I missed before.

25 years agoTypo fix.
Peter Wemm [Wed, 14 Oct 1998 10:04:32 +0000 (10:04 +0000)]
Typo fix.

25 years agoInitial attempt to update the Alpha loader and kernel to use the machine
Peter Wemm [Wed, 14 Oct 1998 09:53:25 +0000 (09:53 +0000)]
Initial attempt to update the Alpha loader and kernel to use the machine
independent elf loader and have access to kld modules.  Jordan and I were
not sure how to create boot floppies, and the things we tried just made
SRM laugh in our faces - but it was upset at boot1 which was not touched
by these changes.  Essentially this has been untested. :-(

What this does is to steal the last three slots from the nine spare longs
in the bootinfo_v1 struct to pass the module base pointer through.

The startup code now to set up and fills in the module and environment
structures, hopefully close enough to the i386 layout to be able to use
the same kernel code.  We now pass though the updated end of the kernel
space used, rather than _end. (like the i386).

If this does not work, it needs to be beaten into shape pronto.  Otherwise
it should be backed out before 3.0.

Pre-approved in principle by: dfr

25 years agorough 1st attempt at cleaning this up. Those feeling brave/knowledgable about
Jordan K. Hubbard [Wed, 14 Oct 1998 08:54:03 +0000 (08:54 +0000)]
rough 1st attempt at cleaning this up.  Those feeling brave/knowledgable about
upgrade issues should also check the section created for this.

25 years agoRevert part of previous commit. vaddr_t doesn't exist on FreeBSD. This
Peter Wemm [Wed, 14 Oct 1998 08:31:27 +0000 (08:31 +0000)]
Revert part of previous commit.  vaddr_t doesn't exist on FreeBSD.  This
didn't affect the x86 kernel due to #ifdefs.  It broke FreeBSD/Alpha
kernel compiles though.

25 years agoTry and get the sys/* and machine/* includes via relative paths. This
Peter Wemm [Wed, 14 Oct 1998 07:08:16 +0000 (07:08 +0000)]
Try and get the sys/* and machine/* includes via relative paths.  This
saves having to do a 'make includes' after touching any header file for
the boot code.

25 years agoAdd man pages for many of the functions in the CAM library. This covers
Kenneth D. Merry [Wed, 14 Oct 1998 06:20:36 +0000 (06:20 +0000)]
Add man pages for many of the functions in the CAM library.  This covers
most of the open/close routines, and the buffer/cdb parsing routines
derived from the old scsi(3) library.

The cam_cdbparse(3) man page borrows from the old scsi(3) man page, so the
copyright and history section reflect that.

The many scsi_* functions and other functions that are pulled in from the
kernel aren't documented yet, but will be eventually.

25 years agoAlign to sizeof(long) rather than sizeof(int32_t). It needs to be
Peter Wemm [Wed, 14 Oct 1998 05:07:23 +0000 (05:07 +0000)]
Align to sizeof(long) rather than sizeof(int32_t).  It needs to be
long because this code is shared with the alpha.  I hope the alpha can
read 32 bit ints at 32 bit alignment (vs. 64 bit alignment).

25 years agoSupport encapsulating ELF linker sets for KLD modules via gensetdefs.
Peter Wemm [Wed, 14 Oct 1998 04:04:22 +0000 (04:04 +0000)]
Support encapsulating ELF linker sets for KLD modules via gensetdefs.
Support name based dependencies at build time.  This is a hack.
These only affect KLD modules.

25 years ago+gensetdefs
Peter Wemm [Wed, 14 Oct 1998 03:41:25 +0000 (03:41 +0000)]
+gensetdefs

25 years agogensetdefs is a standalone version of the tool that we use to wrap a
Peter Wemm [Wed, 14 Oct 1998 03:41:01 +0000 (03:41 +0000)]
gensetdefs is a standalone version of the tool that we use to wrap a
struct linker_set around the contents of ELF linker sets.  This tool
also generates setdef0.c and setdef1.c for the alpha and i386 rather than
having these duplicated all over the tree too.

This is required for building KLD modules.

25 years agoAdd a description of the "panic: cannot mount root" post-install problem, and
Mike Smith [Wed, 14 Oct 1998 02:44:10 +0000 (02:44 +0000)]
Add a description of the "panic: cannot mount root" post-install problem, and
some ways to work around it.

25 years agoFix more bogons revealed by building alpha boot floppies.
Jordan K. Hubbard [Wed, 14 Oct 1998 02:13:45 +0000 (02:13 +0000)]
Fix more bogons revealed by building alpha boot floppies.

25 years agoFix btx include path.
Robert Nordier [Wed, 14 Oct 1998 01:53:56 +0000 (01:53 +0000)]
Fix btx include path.

25 years agoDon't generate a spam-friendly /etc/hosts by default.
Jordan K. Hubbard [Wed, 14 Oct 1998 01:04:44 +0000 (01:04 +0000)]
Don't generate a spam-friendly /etc/hosts by default.
Suggested by: jmb

25 years agoDB_ELF_SYMBOLS doesn't appear to do anything yet, or if it does it's
Jordan K. Hubbard [Wed, 14 Oct 1998 00:42:02 +0000 (00:42 +0000)]
DB_ELF_SYMBOLS doesn't appear to do anything yet, or if it does it's
not been added to the appropriate options file.  Comment it out to
prevent config(8) warnings.

25 years agoTweak the output one more time again. The kernel or module pathname
Peter Wemm [Wed, 14 Oct 1998 00:41:17 +0000 (00:41 +0000)]
Tweak the output one more time again.  The kernel or module pathname
is useful, and usually fits all on one line with the load sizes.

25 years agoInclude <bsd.prog.mk>. Add install target (to /boot for now).
Robert Nordier [Wed, 14 Oct 1998 00:24:16 +0000 (00:24 +0000)]
Include <bsd.prog.mk>.  Add install target (to /boot for now).

25 years agoFix flow of control after directory listing; enable EDD support;
Robert Nordier [Tue, 13 Oct 1998 23:43:38 +0000 (23:43 +0000)]
Fix flow of control after directory listing; enable EDD support;
cosmetics.

25 years agoDisable cache syncs for a broken NEC drive.
Kenneth D. Merry [Tue, 13 Oct 1998 23:34:54 +0000 (23:34 +0000)]
Disable cache syncs for a broken NEC drive.

Reviewed by: gibbs
Submitted by: Blaz Zupan <blaz@gold.amis.net>

25 years agoMake v86.ctl default more explicit; simplify read error-handling;
Robert Nordier [Tue, 13 Oct 1998 23:00:47 +0000 (23:00 +0000)]
Make v86.ctl default more explicit; simplify read error-handling;
twiddle.

25 years agoThe adjtime man page is in section 2, not 8.
Dag-Erling Smørgrav [Tue, 13 Oct 1998 22:55:29 +0000 (22:55 +0000)]
The adjtime man page is in section 2, not 8.

PR: docs/8310
Submitted by: Amakawa Shuhei <amakawa@nebula.sf.t.u-tokyo.ac.jp>

25 years agoAdjust NDEV value.
Robert Nordier [Tue, 13 Oct 1998 22:17:05 +0000 (22:17 +0000)]
Adjust NDEV value.
Optimize reading of system time.

25 years agoWhen we send a stop unit command to a device, send it as an ordered tag so
Kenneth D. Merry [Tue, 13 Oct 1998 22:02:38 +0000 (22:02 +0000)]
When we send a stop unit command to a device, send it as an ordered tag so
that any transactions in front of the stop command get flushed to disk
first.  This will have no effect on devices that have tagged queueing
turned off, or don't support tagged queueing.

Reviewed by: gibbs

25 years agoFix a bug in the error recovery code. It was possible to have more than
Kenneth D. Merry [Tue, 13 Oct 1998 21:41:32 +0000 (21:41 +0000)]
Fix a bug in the error recovery code.  It was possible to have more than
one error recovery action oustanding for a given peripheral.

This is bad for several reasons.  The first problem is that the error
recovery actions would likely be to fix the same problem.  (e.g., we
queue 5 CCBs to a disk, and the first one comes back with 0x04,0x02.  We
start error recovery, and the second one comes back with the same status.
Then the third one comes back, and so on.  Each one causes the drive to get
nailed with a start unit, when we really only need one.)

The other problem is that we only have space to store one CCB while we're
doing error recovery.  The subsequent error recovery actions that got
started were over-writing the CCBs from previous error recovery actions,
but we still tried to call the done routine N times for N error recovery
actions.  Each call to dadone() was done with the same CCB, though.  So on
the second one, we got a "biodone: buffer not busy" panic, since the buffer
in question had already been through biodone().

In any case, this fixes things so that any any given time, there's only one
error recovery action outstanding for any given peripheral driver.

Reviewed by: gibbs
Reported by: Philippe Regnauld <regnauld@deepo.prosa.dk>
[ Philippe wins the "bug finder of the week" award ]

25 years agoSync up with some needed x86 options.
Jordan K. Hubbard [Tue, 13 Oct 1998 21:38:46 +0000 (21:38 +0000)]
Sync up with some needed x86 options.

25 years agowhoops, correct patch-o in last commit.
Jordan K. Hubbard [Tue, 13 Oct 1998 21:37:30 +0000 (21:37 +0000)]
whoops, correct patch-o in last commit.

25 years agoMore alpha patches.
Jordan K. Hubbard [Tue, 13 Oct 1998 21:36:13 +0000 (21:36 +0000)]
More alpha patches.

25 years agoChange to a 15-sector boot2.
Robert Nordier [Tue, 13 Oct 1998 21:35:42 +0000 (21:35 +0000)]
Change to a 15-sector boot2.
Refine slice-handling.

25 years agoFix a bug in the scan lun code that showed up when we did the following
Kenneth D. Merry [Tue, 13 Oct 1998 21:29:04 +0000 (21:29 +0000)]
Fix a bug in the scan lun code that showed up when we did the following
sequence of things:

- spin up a disk
  - send an async event to refresh the inquiry data
    - run through xpt_scan_lun() to re-probe the device
        - eventually finish the probe, but panic in xpt_done() because the
          periph pointer wasn't set.

Reviewed by: gibbs
Reported by: Philippe Regnauld <regnauld@deepo.prosa.dk>

25 years agoProperly conditionalize more stuff for the ALPHA.
Jordan K. Hubbard [Tue, 13 Oct 1998 21:09:42 +0000 (21:09 +0000)]
Properly conditionalize more stuff for the ALPHA.

25 years agoUpdates for Alpha.
Jordan K. Hubbard [Tue, 13 Oct 1998 21:03:56 +0000 (21:03 +0000)]
Updates for Alpha.

25 years agoSet the user context correctly so that cd ~ does the right thing.
Dag-Erling Smørgrav [Tue, 13 Oct 1998 20:42:01 +0000 (20:42 +0000)]
Set the user context correctly so that cd ~ does the right thing.

PR: bin/7943 bin/8293
Submitted by: Bill Fenner <fenner@parc.xerox.com>
Approved by: jkh

25 years agoMake Australian time GB-like
Andrey A. Chernov [Tue, 13 Oct 1998 20:20:27 +0000 (20:20 +0000)]
Make Australian time GB-like
PR: 8302
Submitted by: Stephen McKay <syssgm@dtir.qld.gov.au>

25 years agoDon't use an absolute path to objcopy.
Robert Nordier [Tue, 13 Oct 1998 18:29:18 +0000 (18:29 +0000)]
Don't use an absolute path to objcopy.
Noticed by: Scott Mace <smace@intt.org>

25 years agoImprove drive recognition and handling.
Robert Nordier [Tue, 13 Oct 1998 17:41:06 +0000 (17:41 +0000)]
Improve drive recognition and handling.

25 years agoUpdated library order and comments about it. This fixes libm not
Bruce Evans [Tue, 13 Oct 1998 16:48:43 +0000 (16:48 +0000)]
Updated library order and comments about it.  This fixes libm not
being built before libf2c and libmd not being built before libatm.

25 years agoFix a bug in camcontrol that caused 'camcontrol start' to not work.
Kenneth D. Merry [Tue, 13 Oct 1998 16:23:26 +0000 (16:23 +0000)]
Fix a bug in camcontrol that caused 'camcontrol start' to not work.

Noticed by: Philippe Regnauld <regnauld@deepo.prosa.dk>

25 years agoFixed bloatage of `struct inode'. We used 5 "spare" fields for ext2fs,
Bruce Evans [Tue, 13 Oct 1998 15:45:43 +0000 (15:45 +0000)]
Fixed bloatage of `struct inode'.  We used 5 "spare" fields for ext2fs,
but when i_effnlink was added to support soft updates, there was only
room for 4 spares.  The number of spares was not reduced, so the inode
size became 260 (on i386's), or 512 after rounding up by malloc().
Use one spare field in `struct dinode' instead of the 5th spare field
in the inode and reduced to 4 spares in the inode so that the size is
256 again.

Changed the types of the spares in the inode from int to u_int32_t
so that the inode size has more chance of being <= 256 under other
arches, and downdated ext2fs to match (it was broken to use ints
before rev.1.1).

25 years agoOne vfork() changed to fork(); one execl() changed to execv(). The
Dag-Erling Smørgrav [Tue, 13 Oct 1998 15:14:37 +0000 (15:14 +0000)]
One vfork() changed to fork(); one execl() changed to execv(). The
latter isn't actually called as far as I can see since FreeBSD uses
termcap and not terminfo.

25 years agoMissed one in the previous commit.
Dag-Erling Smørgrav [Tue, 13 Oct 1998 15:05:00 +0000 (15:05 +0000)]
Missed one in the previous commit.

25 years agoCalls one or more of malloc(), warn(), err(), syslog(), execlp() or
Dag-Erling Smørgrav [Tue, 13 Oct 1998 14:52:33 +0000 (14:52 +0000)]
Calls one or more of malloc(), warn(), err(), syslog(), execlp() or
execvp() in the child branch of a vfork(). Changed to use fork()
instead.

Some of these (mv, find, apply, xargs) might benefit greatly from
being rewritten to use vfork() properly.

PR: Loosely related to bin/8252
Approved by: jkh and bde

25 years agoDoh. What was I smoking when I wrote this?
Dag-Erling Smørgrav [Tue, 13 Oct 1998 12:19:31 +0000 (12:19 +0000)]
Doh. What was I smoking when I wrote this?

PR: bin/8301
Submitted by: Tetsuya Furukawa <tetsuya@secom-sis.co.jp>

25 years agofix brainos in earlier commit.
Jordan K. Hubbard [Tue, 13 Oct 1998 10:07:43 +0000 (10:07 +0000)]
fix brainos in earlier commit.

25 years agoInclude btxld in build-tools only for i386.
Robert Nordier [Tue, 13 Oct 1998 10:01:32 +0000 (10:01 +0000)]
Include btxld in build-tools only for i386.

25 years agoproperly conditionalize for alpha.
Jordan K. Hubbard [Tue, 13 Oct 1998 09:49:16 +0000 (09:49 +0000)]
properly conditionalize for alpha.

25 years agocorrect typo.
Jordan K. Hubbard [Tue, 13 Oct 1998 09:46:40 +0000 (09:46 +0000)]
correct typo.

25 years agoswitch to proper boot0
Jordan K. Hubbard [Tue, 13 Oct 1998 09:45:59 +0000 (09:45 +0000)]
switch to proper boot0