]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
25 years agoSync with sys/i386/isa/clock.c revision 1.133.
kato [Mon, 10 May 1999 09:05:01 +0000 (09:05 +0000)]
Sync with sys/i386/isa/clock.c revision 1.133.

25 years agoSync with sys/i386/conf/options.i386 revision 1.113.
kato [Mon, 10 May 1999 09:03:10 +0000 (09:03 +0000)]
Sync with sys/i386/conf/options.i386 revision 1.113.

25 years agoSync with sys/i386/conf/majors.i386 revision 1.74.
kato [Mon, 10 May 1999 09:02:23 +0000 (09:02 +0000)]
Sync with sys/i386/conf/majors.i386 revision 1.74.

25 years agoSync with sys/i386/conf/files.i386 revision 1.239.
kato [Mon, 10 May 1999 09:01:19 +0000 (09:01 +0000)]
Sync with sys/i386/conf/files.i386 revision 1.239.

25 years agoSync with sys/i386/conf/GENERIC and Makefile.i386 revisions 1.169 and
kato [Mon, 10 May 1999 09:00:29 +0000 (09:00 +0000)]
Sync with sys/i386/conf/GENERIC and Makefile.i386 revisions 1.169 and
1.152, respectively.

25 years agoAdd update on the status of new bus.
imp [Mon, 10 May 1999 05:34:32 +0000 (05:34 +0000)]
Add update on the status of new bus.
Add libcam ABI change warning.

25 years agoFix mknod which overloaded a little too much on "minor" and "major"
phk [Mon, 10 May 1999 04:50:27 +0000 (04:50 +0000)]
Fix mknod which overloaded a little too much on "minor" and "major"

25 years agoFixed disordering in previous commit.
bde [Mon, 10 May 1999 03:51:07 +0000 (03:51 +0000)]
Fixed disordering in previous commit.

25 years agoMake corresponding changes to match config update
peter [Mon, 10 May 1999 03:02:22 +0000 (03:02 +0000)]
Make corresponding changes to match config update

25 years agoApparently we need to define rootdev and dumpdev here now. This is
peter [Mon, 10 May 1999 02:56:38 +0000 (02:56 +0000)]
Apparently we need to define rootdev and dumpdev here now.  This is
(obviously) not finished yet.

25 years agoUnderstand 'config kernelname' and pretend it was a
peter [Mon, 10 May 1999 02:42:58 +0000 (02:42 +0000)]
Understand 'config kernelname' and pretend it was a
'makeoptions KERNEL=kernelname'.  Warn about any trailing stuff as it's
not handled here.  This is a simple bandaid, hopefully to head off some
complaints from certain people.

25 years agoMFS.
obrien [Mon, 10 May 1999 00:54:18 +0000 (00:54 +0000)]
MFS.

25 years agoSometime since this file was written, the list of kernel malloc types
imp [Mon, 10 May 1999 00:33:32 +0000 (00:33 +0000)]
Sometime since this file was written, the list of kernel malloc types
changed from a simple list to a circular one.  We compensate by only
looping until we see the first address again.  Before, things would
terminate because it was limited to 200 iterations.  This lead to
bogus statistics and repeating stats for memory types.

This should be merged into 3.2, as the same bug is there.

25 years ago#if 0 some unused code (debug?) to quiet a warning.
peter [Mon, 10 May 1999 00:20:46 +0000 (00:20 +0000)]
#if 0 some unused code (debug?) to quiet a warning.

25 years agoClean out some unused leftovers from before the split from the old isa.c.
peter [Sun, 9 May 1999 23:56:00 +0000 (23:56 +0000)]
Clean out some unused leftovers from before the split from the old isa.c.

25 years agoUn-duplicate DO_SETBITS() (it was identical for both SMP and !SMP)..
peter [Sun, 9 May 1999 23:40:29 +0000 (23:40 +0000)]
Un-duplicate DO_SETBITS() (it was identical for both SMP and !SMP)..

25 years agoFor what it's worth, idelayed is declared as a volatile in the headers,
peter [Sun, 9 May 1999 23:32:29 +0000 (23:32 +0000)]
For what it's worth, idelayed is declared as a volatile in the headers,
and even though it's not used in this file make it a volatile here too.

25 years agoloadandclear() uses an atomic instruction (even on SMP, where it's an
peter [Sun, 9 May 1999 23:30:01 +0000 (23:30 +0000)]
loadandclear() uses an atomic instruction (even on SMP, where it's an
implicitly LOCK'ed instruction), so there shouldn't be any harm in making
it volatile pointer compatable for one of the users of it.  It seems to
generate the same code regardless.

25 years ago#ifdef BRIDGE around a goto label used by the bridge code to silcence a
peter [Sun, 9 May 1999 23:24:47 +0000 (23:24 +0000)]
#ifdef BRIDGE around a goto label used by the bridge code to silcence a
warning.

25 years agoReally support Ultra-2 chips. Symbios redefined a register that
se [Sun, 9 May 1999 22:44:42 +0000 (22:44 +0000)]
Really support Ultra-2 chips. Symbios redefined a register that
was available to the programmer to hold chip state information:

Use the SDID register instead of CTEST3. This change actually
simplifies the SCRIPTS code, but I'm not absolutely sure, that
it is OK for all variants of NCR chips around and all device
combinations. I have had this code running on several systems
with 53c810, 875 and 895 controllers for several months.

Suggested by: Gerard Roudier <groudier@club-internet.fr>

25 years agoMove KERNEL?=kernel to top of the file where it's more obvious and add
peter [Sun, 9 May 1999 22:42:17 +0000 (22:42 +0000)]
Move KERNEL?=kernel to top of the file where it's more obvious and add
a pointer to 'makeoptions' and /etc/make.conf.
Catch a few stray "kernel" hardcoded references.
Move the kernel.debug and related kernel build rules together.

25 years agoPut an example of 'makeoptions KERNEL=foo' to replace the old
peter [Sun, 9 May 1999 22:26:10 +0000 (22:26 +0000)]
Put an example of 'makeoptions KERNEL=foo' to replace the old
'config foo' functionality.

25 years agoSilence message about absence of PnP overrides.
abial [Sun, 9 May 1999 22:16:27 +0000 (22:16 +0000)]
Silence message about absence of PnP overrides.

25 years agoRemove hackish foot protection.. If you ask for an atkbd on an isa bus
peter [Sun, 9 May 1999 20:45:53 +0000 (20:45 +0000)]
Remove hackish foot protection..  If you ask for an atkbd on an isa bus
rather than on the keyboard controller, you will get it.. (and shoot your
feet)

25 years agoLites2 seems to have pretty much disappeared from the radar, and I suspect
peter [Sun, 9 May 1999 20:42:45 +0000 (20:42 +0000)]
Lites2 seems to have pretty much disappeared from the radar, and I suspect
far more than this hack would be needed now..

25 years agoMake COM_ESP and COM_MULTIPORT compile for LINT. It might even work..
peter [Sun, 9 May 1999 20:35:02 +0000 (20:35 +0000)]
Make COM_ESP and COM_MULTIPORT compile for LINT.  It might even work..

25 years agodon't confuse units and devices.
phk [Sun, 9 May 1999 20:29:04 +0000 (20:29 +0000)]
don't confuse units and devices.

25 years agoFor the ioctl that reads the pci configuration, look up the name and unit
peter [Sun, 9 May 1999 20:27:26 +0000 (20:27 +0000)]
For the ioctl that reads the pci configuration, look up the name and unit
on the fly so that we can see the driver assignment of new pci devices
as well in the 'pciconf -l' display.

25 years agoGC unused variable in struct.
peter [Sun, 9 May 1999 20:25:01 +0000 (20:25 +0000)]
GC unused variable in struct.

25 years agoRevert to the previous behaviour of only doing a utmp
brian [Sun, 9 May 1999 20:13:52 +0000 (20:13 +0000)]
Revert to the previous behaviour of only doing a utmp
login if we're -direct on a tty device.

25 years agoEnable the support for a file interface on the HID driver for
n_hibma [Sun, 9 May 1999 20:13:51 +0000 (20:13 +0000)]
Enable the support for a file interface on the HID driver for
easy testing of the HID reports that come back.

Submitted by: MAEKAWA Masahide <bishop@rr.iij4u.or.jp>

25 years agoDeal with the fact that as we now mbuf_Read the fsm
brian [Sun, 9 May 1999 20:02:29 +0000 (20:02 +0000)]
Deal with the fact that as we now mbuf_Read the fsm
header in fsm_Input() we often end up with a NULL mbuf.

Deal with a possible NULL mbuf being passed into
mbuf_Prepend().

Adjust some spacing to make things more consistent.

25 years agoPut back changes that might be causing trouble on Alpha.
mckusick [Sun, 9 May 1999 19:39:54 +0000 (19:39 +0000)]
Put back changes that might be causing trouble on Alpha.

25 years agoReserve 124 for USB Communications Class Drivers
n_hibma [Sun, 9 May 1999 19:38:19 +0000 (19:38 +0000)]
Reserve 124 for USB Communications Class Drivers

25 years agoAdd the USB Mass Storage manual page
n_hibma [Sun, 9 May 1999 19:35:51 +0000 (19:35 +0000)]
Add the USB Mass Storage manual page

25 years agos/main/mi_startup/
peter [Sun, 9 May 1999 19:03:55 +0000 (19:03 +0000)]
s/main/mi_startup/

25 years agos/main/mi_startup/ for the kernel entry point so that egcs doesn't get
peter [Sun, 9 May 1999 19:01:49 +0000 (19:01 +0000)]
s/main/mi_startup/ for the kernel entry point so that egcs doesn't get
upset about it (and generate things like __main() calls that are reserved
for main()).  Renaming was phk's suggestion, but I'd already thought about
it too.  (phk liked my suggested name tada() but I decided against it :-)

Reviewed by: phk

25 years agoMove a couple of bits of (now) static content from config(8) (ie: %LOAD)
peter [Sun, 9 May 1999 18:57:02 +0000 (18:57 +0000)]
Move a couple of bits of (now) static content from config(8) (ie: %LOAD)
directly into the Makefile.  Remove references to swapkernel.c, it's
not generated by config(8) now.  (The previous config commits had
generated it, but they had an unused 'char *' in them).

25 years agoPut on my viking helmet from the closet, and get out the war axe and
peter [Sun, 9 May 1999 18:54:25 +0000 (18:54 +0000)]
Put on my viking helmet from the closet, and get out the war axe and
"retire" some more stuff.

25 years agoAdded Thrustmaster Fusion Digital Gamepad
n_hibma [Sun, 9 May 1999 18:29:37 +0000 (18:29 +0000)]
Added Thrustmaster Fusion Digital Gamepad

25 years agoMake minor, major and makedev inline functions so we gain type checking
phk [Sun, 9 May 1999 18:03:48 +0000 (18:03 +0000)]
Make minor, major and makedev inline functions so we gain type checking
and prevent compile time creation of dev_t's from happening again.

25 years agoWe don't have composite devices here either.
peter [Sun, 9 May 1999 17:48:23 +0000 (17:48 +0000)]
We don't have composite devices here either.

25 years agoNo point in swapdev being a static global when used only locally.
phk [Sun, 9 May 1999 17:28:00 +0000 (17:28 +0000)]
No point in swapdev being a static global when used only locally.

25 years agoconfig(8) lobotomy, please see commit msg in sys.
phk [Sun, 9 May 1999 17:23:38 +0000 (17:23 +0000)]
config(8) lobotomy, please see commit msg in sys.

(I have no idea why cvs didn't take these changes before.)

25 years agoMissing revision markers.
peter [Sun, 9 May 1999 17:12:25 +0000 (17:12 +0000)]
Missing revision markers.

25 years agoSimplify the COMPAT_PCI_DRIVER/DATA_SET hack. We can add:
peter [Sun, 9 May 1999 17:07:30 +0000 (17:07 +0000)]
Simplify the COMPAT_PCI_DRIVER/DATA_SET hack.  We can add:
#define COMPAT_PCI_DRIVER(name,data) DATA_SET(pcidevice_set,data)
.. to 2.2.x and 3.x if people think it's worth it.  Driver writers can do
this if it's not defined.  (The reason for this is that I'm trying to
progressively eliminate use of linker_sets where it hurts modularity and
runtime load capability, and these DATA_SET's keep getting in the way.)

25 years agoTurn on compat22 dist.
obrien [Sun, 9 May 1999 17:00:04 +0000 (17:00 +0000)]
Turn on compat22 dist.

25 years agoOptimize out a couple of places where com_addr() is used twice in a row,
peter [Sun, 9 May 1999 16:56:38 +0000 (16:56 +0000)]
Optimize out a couple of places where com_addr() is used twice in a row,
although this is pretty trivial.  devclass_get_softc() is a tad more
expensive than the old com_addr() implementation.  If Bruce is really
worried about the cost of this, it could always be changed so that the
softc pointer is stored in a local array again.

25 years agoOops. If ROOTDEVNAME isn't defined, have -r call -a.
phk [Sun, 9 May 1999 16:56:09 +0000 (16:56 +0000)]
Oops.  If ROOTDEVNAME isn't defined, have -r call -a.

25 years agono longer used.
phk [Sun, 9 May 1999 16:46:38 +0000 (16:46 +0000)]
no longer used.

25 years agoMajor lobotomy of config(8). The
phk [Sun, 9 May 1999 16:46:01 +0000 (16:46 +0000)]
Major lobotomy of config(8).  The

        config kernel mumble mumble

line has been obsoleted and removed and with it went all knowledge of
devices on the part of config.

You can still configure a root device (which is used if you give
the "-r" flag) but now with an option:

        options         ROOTDEVNAME=\"da0s2e\"

The string is parsed by the same code as at the "boot -a" prompt.

At the same time, make the "boot -a" prompt both more able and more
informative.

ALPHA/PC98 people:  You will have to adapt a few simple changes
(defining rootdev and dumpdev somewhere else) before config works
for you again, sorry, but it's all in the name of progress.

25 years agoDelete the #if'ed out old isa bus probe/attach routines
peter [Sun, 9 May 1999 16:39:24 +0000 (16:39 +0000)]
Delete the #if'ed out old isa bus probe/attach routines

25 years agoMake the pci vga detection work so that the card is claimed and it's
peter [Sun, 9 May 1999 16:32:00 +0000 (16:32 +0000)]
Make the pci vga detection work so that the card is claimed and it's
interrupt configuration reported.  (I just discovered my vga card is
being configured for irq 5... :-)  This is just reporting.  The vga_isa
driver does the real work using the isa compat mappings.

25 years agoFix a couple of warnings and some bitrot in comments.
peter [Sun, 9 May 1999 16:04:14 +0000 (16:04 +0000)]
Fix a couple of warnings and some bitrot in comments.

25 years agoAlways use the module system, rather than in an #ifdef.
peter [Sun, 9 May 1999 15:57:52 +0000 (15:57 +0000)]
Always use the module system, rather than in an #ifdef.

25 years agoArgh, don't clobber the pci device list if there are multiple busses!
peter [Sun, 9 May 1999 15:54:04 +0000 (15:54 +0000)]
Argh, don't clobber the pci device list if there are multiple busses!
(An AGP counts as a PCI bus, it seems...)
This stopped 'pciconf -l' from working on AGP or PCI->PCI bridge systems.

25 years agoMore typos.
kris [Sun, 9 May 1999 13:46:31 +0000 (13:46 +0000)]
More typos.

25 years agoTypo.
kris [Sun, 9 May 1999 13:41:19 +0000 (13:41 +0000)]
Typo.

Obtained from: OpenBSD

25 years agoQuiet a warning, put a #if 0 around some (I believe) unused static structs.
peter [Sun, 9 May 1999 13:21:06 +0000 (13:21 +0000)]
Quiet a warning, put a #if 0 around some (I believe) unused static structs.

25 years agomajor(something) can never become NODEV.
phk [Sun, 9 May 1999 13:13:52 +0000 (13:13 +0000)]
major(something) can never become NODEV.

25 years agoremove cast from dev_t to dev_t.
phk [Sun, 9 May 1999 13:11:37 +0000 (13:11 +0000)]
remove cast from dev_t to dev_t.

25 years agoMove opt_sio.h options to conf/options and activate in isa/sio.c
peter [Sun, 9 May 1999 13:10:49 +0000 (13:10 +0000)]
Move opt_sio.h options to conf/options and activate in isa/sio.c

25 years agoUnconfuse DEV_MODULE() and DEV_DRIVER_MODULE() about the difference between
phk [Sun, 9 May 1999 13:00:50 +0000 (13:00 +0000)]
Unconfuse DEV_MODULE() and DEV_DRIVER_MODULE() about the difference between
a major number for a dev_t.

25 years agoRemove DDB.
dfr [Sun, 9 May 1999 11:42:19 +0000 (11:42 +0000)]
Remove DDB.

25 years agoBack out the last change to disk.c - a kernel change removed the need.
dfr [Sun, 9 May 1999 11:34:58 +0000 (11:34 +0000)]
Back out the last change to disk.c - a kernel change removed the need.
Enabled writing to the disklabel when writing a disk (only affects alpha).

25 years agoHack the diskslice stuff so that it allows the alpha sysinstall to
dfr [Sun, 9 May 1999 11:27:41 +0000 (11:27 +0000)]
Hack the diskslice stuff so that it allows the alpha sysinstall to
manipulate the disklabel. This is almost certainly not the right way
to do it but I'm desperate.

25 years agoadd some amount of sanity to the way the gdb stuff finds its device.
phk [Sun, 9 May 1999 10:51:13 +0000 (10:51 +0000)]
add some amount of sanity to the way the gdb stuff finds its device.

I'm not too happy about the result either, but at least it has less
chance of backfiring.

This particular feature could be called "a mess" without offending
anybody.

25 years agoFix two warnings.
peter [Sun, 9 May 1999 10:45:54 +0000 (10:45 +0000)]
Fix two warnings.

25 years agoFix warnings.
peter [Sun, 9 May 1999 10:43:54 +0000 (10:43 +0000)]
Fix warnings.

25 years agoYet a major/dev_t confusion
phk [Sun, 9 May 1999 10:42:39 +0000 (10:42 +0000)]
Yet a major/dev_t confusion

Spotted by Bruce.

25 years agosiocngetc & putc can be staticized.
phk [Sun, 9 May 1999 10:28:50 +0000 (10:28 +0000)]
siocngetc & putc can be staticized.

25 years agoYet another place which knew too much. Still not sure how much
phk [Sun, 9 May 1999 10:25:30 +0000 (10:25 +0000)]
Yet another place which knew too much.  Still not sure how much
good this does in the end.

25 years agofix some DRIVER_TYPE / INTR_TYPE confusions.
phk [Sun, 9 May 1999 09:56:52 +0000 (09:56 +0000)]
fix some DRIVER_TYPE / INTR_TYPE confusions.

25 years agoDon't confuse dev_t and major numbers in DEV_MODULE()
phk [Sun, 9 May 1999 08:58:45 +0000 (08:58 +0000)]
Don't confuse dev_t and major numbers in DEV_MODULE()

25 years agoPeter beat me to half this patch, but didn't do the other half:
phk [Sun, 9 May 1999 08:18:12 +0000 (08:18 +0000)]
Peter beat me to half this patch, but didn't do the other half:

set d_bmaj

don't cast a dev_t to int before comparing to NODEV

25 years agoComment advising ordering of cdevsw_add and bdevsw_add is obsolete (no
peter [Sun, 9 May 1999 08:10:17 +0000 (08:10 +0000)]
Comment advising ordering of cdevsw_add and bdevsw_add is obsolete (no
bdevsw_add any more).

25 years agoDuh, bdevsw() takes dev_t arg.
phk [Sun, 9 May 1999 07:56:36 +0000 (07:56 +0000)]
Duh, bdevsw() takes dev_t arg.

25 years agoMutex test program.
jb [Sun, 9 May 1999 07:55:24 +0000 (07:55 +0000)]
Mutex test program.

Submitted by: Daniel Eischen <eischen@vigrid.com>

25 years agoAdd mutex test program.
jb [Sun, 9 May 1999 07:50:31 +0000 (07:50 +0000)]
Add mutex test program.

25 years agoRemoved DRIVER_TYPE_TTY field.
kato [Sun, 9 May 1999 05:02:26 +0000 (05:02 +0000)]
Removed DRIVER_TYPE_TTY field.

25 years agoReserve a major number for the frame buffer device.
yokota [Sun, 9 May 1999 05:02:15 +0000 (05:02 +0000)]
Reserve a major number for the frame buffer device.

25 years agoSync with sys/isa/sio.c revision 1.234.
kato [Sun, 9 May 1999 05:00:54 +0000 (05:00 +0000)]
Sync with sys/isa/sio.c revision 1.234.

25 years agoDon't confuse cursor keys with numpad keys when composing a char code.
yokota [Sun, 9 May 1999 05:00:19 +0000 (05:00 +0000)]
Don't confuse cursor keys with numpad keys when composing a char code.

PR: kern/10988

25 years agoMinor tweak after the introduction of new-bus to i386; properly
yokota [Sun, 9 May 1999 04:59:24 +0000 (04:59 +0000)]
Minor tweak after the introduction of new-bus to i386; properly
check "disabled" and "flags" probe hints.

25 years agoRevive APM hooking code for i386 arch now that the same source file is
yokota [Sun, 9 May 1999 04:58:35 +0000 (04:58 +0000)]
Revive APM hooking code for i386 arch now that the same source file is
used for both i386 and alpha (the code was in sys/i386/isa/psm.c, but
was disabled when ported to alpha.)

25 years agoMake apm_probe() properly return an error code when APM BIOS calls
yokota [Sun, 9 May 1999 04:58:13 +0000 (04:58 +0000)]
Make apm_probe() properly return an error code when APM BIOS calls
failed, so that the apm driver won't be attached.

25 years agoAdded backward compatibility to set key repeat rate.
yokota [Sun, 9 May 1999 04:57:51 +0000 (04:57 +0000)]
Added backward compatibility to set key repeat rate.

Requested by: bde

25 years agoRemoved DRIVER_TYPE_TTY field.
kato [Sun, 9 May 1999 04:56:42 +0000 (04:56 +0000)]
Removed DRIVER_TYPE_TTY field.

25 years agoSync with sys/isa/atkbd_isa.c revision 1.4.
kato [Sun, 9 May 1999 04:53:58 +0000 (04:53 +0000)]
Sync with sys/isa/atkbd_isa.c revision 1.4.

25 years agoSync with sys/i386/isa/wd.c revision 1.197.
kato [Sun, 9 May 1999 04:42:10 +0000 (04:42 +0000)]
Sync with sys/i386/isa/wd.c revision 1.197.

25 years agoSync with sys/i386/isa/pcausio.c revision 1.48.
kato [Sun, 9 May 1999 04:39:36 +0000 (04:39 +0000)]
Sync with sys/i386/isa/pcausio.c revision 1.48.

25 years agoSync with sys/i386/isa/npx.c revision 1.61.
kato [Sun, 9 May 1999 04:38:27 +0000 (04:38 +0000)]
Sync with sys/i386/isa/npx.c revision 1.61.

25 years agoSync with sys/i386/isa/mse.c revision 1.42.
kato [Sun, 9 May 1999 04:37:17 +0000 (04:37 +0000)]
Sync with sys/i386/isa/mse.c revision 1.42.

25 years agoSync with sys/i386/isa/fd.c revision 1.142.
kato [Sun, 9 May 1999 04:34:28 +0000 (04:34 +0000)]
Sync with sys/i386/isa/fd.c revision 1.142.

25 years agoSync with sys/i386/i386/userconfig.c revision 1.141.
kato [Sun, 9 May 1999 04:30:08 +0000 (04:30 +0000)]
Sync with sys/i386/i386/userconfig.c revision 1.141.

25 years agoSync with sys/i386/conf/majors.i386 revision 1.72.
kato [Sun, 9 May 1999 04:28:23 +0000 (04:28 +0000)]
Sync with sys/i386/conf/majors.i386 revision 1.72.

25 years agoSync with sys/i386/conf/files.i386 revision 1.238.
kato [Sun, 9 May 1999 04:27:30 +0000 (04:27 +0000)]
Sync with sys/i386/conf/files.i386 revision 1.238.

25 years agoSync with sys/i386/conf/Makefile.i386 revision 1.149.
kato [Sun, 9 May 1999 04:26:06 +0000 (04:26 +0000)]
Sync with sys/i386/conf/Makefile.i386 revision 1.149.

25 years agoAdd a facility in the CAM error handling code to retry selection timeouts.
ken [Sun, 9 May 1999 01:25:34 +0000 (01:25 +0000)]
Add a facility in the CAM error handling code to retry selection timeouts.
If the client requests that the error recovery code retry a selection
timeout, it will be retried after half a second.  The delay is to give the
device time to recover.

For most of these drivers, I only added selection timeout retries where
they were also retrying unit attention type errors.  The sa(4) driver calls
saerror() in a number of places, but most of them don't request retrying
unit attentions.

Also, bump the default minimum CD changer timeout from 2 to 5 seconds and
the maximum timeout from 10 to 15 seconds.  Some Pioneer changers seem to
have trouble with the shorter timeout.

Reviewed by: gibbs

25 years agoFix a freelist trashing under following confitions:
dt [Sat, 8 May 1999 22:46:46 +0000 (22:46 +0000)]
Fix a freelist trashing under following confitions:
- first program lock a region in a file,
- second program wait on the lock,
- first program extend the region,
- second program interrupted by a signal.