Sheldon Hearn [Thu, 22 Nov 2001 14:26:30 +0000 (14:26 +0000)]
Add the script that will be used to maintain src/share/misc/pci_vendors.
The script written and used originally by msmith has been lost.
This version takes the Boemler and Heckenbach lists and produces merged
output. It defaults to ignoring any entries from Heckenbach already
found in Boemler but the -l option causes it to take the entry with the
longest description where an entry appears in both lists.
If this script is replaced, care should be taken to
1) Always use upper-case hexidecimal tokens in device ids.
2) Always keep device lists sorted within vendor lists, which must also
be sorted.
3) Do not try to include input from the previous pci_vendors file, since
bogus ids seem to be removed from both the Boemler and Heckenbach
lists from time to time.
Benno Rice [Thu, 22 Nov 2001 05:08:35 +0000 (05:08 +0000)]
Change the failure mode in option parsing to silently bailing out of option
negotiation rather than rejecting the request.
Apple OpenFirmware 3.0f3 (the version in my iMac) adds trailing garbage to the
end of an otherwise valid request. Without this change, the requests were
rejected which prevented me from booting.
John Baldwin [Thu, 22 Nov 2001 04:02:09 +0000 (04:02 +0000)]
Switch to using a standard /boot directory including the GENERIC kernel
on CD's 1 and 2. Create a simple /boot/loader.conf on each CD that loads
/boot/mfsroot.gz as a mfs_root image. The mfsroot moved from /mfsroot.gz
to /boot/mfsroot.gz so that all of the boot stuff is in /boot. This allows
a full kernel to be used when installing from CD on all architectures
that support CD booting. Note that on x86, this requires using
/boot/cdboot as a no-emulation boot image instead of /floppies/boot.flp.
John Baldwin [Wed, 21 Nov 2001 23:16:26 +0000 (23:16 +0000)]
Bump the CD boot loader up to 1.1 so that the FICL upgrade a while back
will not fail thinking that the loader version is stale. This lets us
use the Forth code on the CD now.
Orion Hodson [Wed, 21 Nov 2001 04:06:52 +0000 (04:06 +0000)]
Set extmode according to extcaps. Fixes codec sample rate failures
that causes noisy playback with rates other than 48k (reported by
Olexander Kunytsa).
John Baldwin [Tue, 20 Nov 2001 20:27:54 +0000 (20:27 +0000)]
Bump the mfsroot up from 2880 to 3200 for Alpha so it fits again. Use
a label of auto instead of the fixed size minimum2 so that the bump in
size actually has an effect.
John Baldwin [Tue, 20 Nov 2001 20:25:12 +0000 (20:25 +0000)]
- Remove the -T from the 4.x newfs as well.
- Don't write out 0xaa55 to the end of the filesystem as md(4) doesn't
generate that warning and it breaks 'disklabel mdX auto'.
- Use the 'canonical' disk name mdX instead of /dev/mdX to disklabel so
we can use disklabel auto for mfsroot filesizes that aren't a standard
size in /etc/disktab.
Akinori MUSHA [Tue, 20 Nov 2001 18:41:01 +0000 (18:41 +0000)]
(null commit to correct the previous log message)
Now that all the printf invocations from within the system startup
scripts *have been replaced with equivalent handmade functions*, we
can safely remove it.
Please steal the technique from src/etc/MAKEDEV when you want to use
printf(1) functionality from within the system rc scripts.
Sheldon Hearn [Tue, 20 Nov 2001 17:38:31 +0000 (17:38 +0000)]
(null commit)
rev 1.24 only makes sense given that test(1) has been made a shell
builtin command. Without this optimization, you'd be better off using
printf(1) if you cared about efficiency at this level (although the use
of printf(1) prior to mounting /usr would remain an issue).
Sheldon Hearn [Tue, 20 Nov 2001 17:33:10 +0000 (17:33 +0000)]
(null commit)
The previous delta only makes sense given that test(1) has been made a
shell builtin command. Without this optimization, you'd be better off
using printf(1) if you cared about efficiency at this level (although
the use of printf(1) prior to mounting /usr would remain an issue).
Sheldon Hearn [Tue, 20 Nov 2001 17:24:35 +0000 (17:24 +0000)]
Introduce new shell functions hexdigit, hexprint and zeropad.
Use these new functions instead of printf(1), which is scheduled for
removal as a shell builtin command, and which will not be available as a
standalone utility if MAKEDEV is run prior to mounting /usr.