[plug] Intel Motherboard allocates incorrect resouces

[Onno Benschop]

AAAAAAAAAAAAAAAAAAArg...

[Deep Breath]

Summary:
	It's still not working and I'm at my whits end.

Extended Version:

Reminder: 
The machine has 2 x hdd on primary channel, and one tape drive on the
secondary channel. I cannot talk to the tape drive or the secondary
channel, even though the BIOS shows all three devices. The machine used
to show two channels when there was one hdd and one cdrom. (I've used
2.4.18 and am using 2.4.21 - neither works)


Using the BIOS ID string (pt84520a.86a.0009.p04.0205291548), I've
identified this machines motherboard as an Intel d845ebg2 board. Lucky
for me it comes with the 845e chipset - which to my delight seems to be
a problem for people the world over - even though it's been on the
market for well over a year now.

The problem is documented along the lines of:
Intel's ICH4 isn't properly supported yet, which results in people not
being able to set DMA modes, resources collisions and unassigned I/O
ports: <for future surfers (hi!)>
      * Region 0: I/O ports at <unassigned>
      * PCI: Ignoring BAR0-3 of IDE controller 00:1f.1
      * PCI: Found IRQ 10 for device 00:1f.1
      * ICH4: IDE controller at PCI slot 00:1f.1
      * PCI: Device 00:1f.1 not available because of resource collisions

Alan Cox responded in a posting in July 2002 to LKML thus:
        If you look with lspci -v you will find your BIOS has mismapped
        or forgotten to map some of the control register space for that
        device.

It is said that his kernel (2.4.19-ac1) supports this chipset, but what
I cannot determine is if this has been incorporated into 2.4.21 (or
later) or not.

If someone could tell me how to find out, that would be a great help.


I've also initated conversation with Intel, who helpfully suggest I
contact my "Linux manufacturer" for support. They suggest that this
mythical entity can supply me with a BIOS flashing utility that runs
under Linux - suffice to say that I pointed out that Intel was the
supplier of the BIOS and the Flashing Utility.

I've had a *little* luck with running 'dosemu' and BIOS tools, but I
cannot characterise my venture as a success. Some (DOS) tools told me
that they didn't run under Windows (I wasn't :-) and others crashed
dosemu. Some actually came back with some answers, but not enough to be
actually useful.

So. While future browsers will come across this message that looks to be
the answer to all their prayers - ha! - I'm still in need of a solution.

Also, the person who suggests that I install a floppy drive or a CD-ROM
gets a gold-star and an elephant stamp, but has to sit in the corner for
the remainder of this adventure :-)

(Idle thought: I wonder if I can net-boot this machine off a remote
device that updates the BIOS.[There is another Debian box sitting next
to it with all the mod-cons.])

Aside: I've disabled all of the peripherals, Audio, On-board LAN,
parallel, serial and it appears as if some of the resource conflicts
have changed IRQ - not enough to actually bring up the second channel
tho.


Anyone got any ideas?

Summary:
     1. How do I determine if AC patches made it into x.y.z kernel
        version?
     2. Is there a Linux BIOS flash tool?
     3. Is there a way to run Intel BIOS Flash tools under Linux?
     4. Does anyone run a kernel with this chipset and BIOS that does
        show the second IDE channel?

[Deep Breath]

Thanks for reading this far... <g>

Onno Benschop 
--

PS. Sorry for any spelling errors. I still cannot get Evolution 1.2.4 to
spell check under Debian testing.

Connected via Optus B3 at S15:51'18" - E128:45'05" (Crossing Falls, Kununurra, WA)
-- 
()/)/)()        ..ASCII for Onno.. 
|>>?            ..EBCDIC for Onno.. 
--- -. -. ---   ..Morse for Onno.. 

Proudly supported by Skipper Trucks, Highway1, Concept AV, Sony Central, Dalcon
ITmaze - ABN: 56 178 057 063 - ph: 04 1219 8888 - onno at itmaze dot com dot au