[coreboot] P2B-LS help: Onboard SCSI BIOS failed to boot

[Keith Hui]

Raised debug level in SeaBIOS. Here is what's new from "Running option
rom at c800:0003" in the old log:

pmm call arg1=0
pmm00: length=1000 handle=29400131 flags=1
pmm_malloc zone=0x000f5770 handle=29400131 size=65536 align=10 ret=0x00080000 ()
pmm call arg1=0
pmm00: length=400 handle=29400132 flags=1
pmm_malloc zone=0x000f5770 handle=29400132 size=16384 align=10 ret=0x0007c000 ()
pnp call arg1=60
pnp call arg1=61
invalid handle_1ab103:78:
   a=0000b103  b=00000000  c=00010000  d=000003bc ds=8000 es=7c00 ss=0000
  si=00000001 di=00005387 bp=00006ab2 sp=00006ab0 cs=8000 ip=9f11  f=0002

This is the Ralf Brown Interrupt List entry for the call being made:

--------X-1AB103-----------------------------

INT 1A - PCI BIOS v2.0c+ - FIND PCI CLASS CODE

	AX = B103h

	ECX = class code (see also #F0085,#00878)

	    bits 31-24 unused

	    bits 23-16 class

	    bits 15-8  subclass

	    bits 7-0   programming interface

	SI = device index (0-n)

Return: CF clear if successful

	CF set on error

	AH = status (00h,86h) (see #00729)

	    00h successful

		BH = bus number

		BL = device/function number (bits 7-3 device, bits 2-0 func)

	    86h device not found

	EAX, EBX, ECX, and EDX may be modified

	all other flags (except IF) may be modified

Notes:	this function may require up to 1024 byte of stack; it will not enable

	  interrupts if they were disabled before making the call

	the meanings of BL and BH on return were exchanged between the initial

	  drafts of the specification and final implementation

	all devices sharing the same Class Code may be enumerated by

	  incrementing SI from 0 until error 86h is returned

SeeAlso: AX=B183h

---

It's looking for the second mass storage device in the system, which
seems perfectly reasonable.
So some PCI bios support seems to be amiss. Something to take to
SeaBIOS mailing list?

Also I'm finding without the option rom added, Linux now was able to
load aic7xxx module and assign an irq to it.

I'm not having fun on the SerialICE front. The Fedora qemu is 0.13.0;
0.14.0 has just been released. I have to download 0.11.0 per
instructions. I'm using lua 5.1.4 as provided by Fedora, because I'm
not sure if my 64bit laptop needs any patch. Website says patch needed
on 32bit platforms; the patch is about 64bit. Then once I have the
SerialICE chip flashed, qemu patched and compiled, lua stuff in place,
then it complains it cannot "index 'regs' because it's nil". So I have
no access to any registers and no idea where things are going because
I can't see CS:EIP.
I'm using the P2B support on P2B-LS. Again I'm trying to figure out
how to switch on/off termination for the onboard SCSI.

I think I am stuck and need more help.

Thanks
Keith

On Sat, Feb 19, 2011 at 8:31 AM, Idwer Vollering <vidwer at gmail.com> wrote:
> 2011/2/19 Keith Hui <buurin at gmail.com>:
>> I am trying (again) to get the Adaptec SCSI on my P2B-LS to initialize
>> properly with coreboot. But it won't boot.
>>
>> Attached is a serial log of what happened.
>>
>> After compiling coreboot I checked out SeaBIOS through git and added
>> the SCSI option rom extracted from my factory bios. A port and a few
>> other resources is assigned, but no IRQ. The SCSI option rom would
>> initialize, I can press the hotkey to enter the utility, but that's
>> it. That option rom hangs before getting any responses from the SCSI
>> hardware.
>
> You want to see what happens during execution.
> Can you create a separate git tree ( git clone
> git://git.seabios.org/seabios.git <dirname_of_choice> ) where you run
> "make menuconfig" in, then set CONFIG_DEBUG_LEVEL to 8 (make
> menuconfig > debugging > debug level) ? Have a look at the other
> options too, maybe.
> Add the resulting file (out/bios.bin.elf) as the payload.
>
>>
>> I am trying to set up SerialICE to see what is going on. Just looking
>> for some insights before I reach that point.
>>
>> I have to jumper off the SCSI to get it to boot. I don't know if this
>> has anything to do but the Ultra2 port on my other P2B-LS is fubared
>> and nothing plugged there would register. Afraid the same may befall
>> this board too.
>
> I have an unused dawicontrol dc-974 (scsi2 ?) that I should try booting off.
>
>>
>> Thanks
>> Keith
>>
>> --
>> coreboot mailing list: coreboot at coreboot.org
>> http://www.coreboot.org/mailman/listinfo/coreboot
>>
>