arima/hdama fails from current cvs

[ron minnich]

I am finally getting time to do K8s again.

I just did a cvs update, built linuxbios for the arima/hdama, and tested 
it out. It failed.

Looks like the ethernet is now back on bus 2, and linux does not see the 
ethernets at all.

Anyone else tried this lately? The K8 platforms seem, overall, to be not 
quite working at present.

ron

[Stefan Reinauer]

* ron minnich  [031104 00:07]:
> I just did a cvs update, built linuxbios for the arima/hdama, and tested 
> it out. It failed.

Any problems besides the bus problem on hdama? It seems that LinuxBIOS
goes some way of emerge and pucker on K8 lately.
> Looks like the ethernet is now back on bus 2, and linux does not see the 
> ethernets at all.

Maybe an additional entry in irq_tables.c helps?
 /* Let Linux know about bus 1 */
 {0x01, (0x04<<3)|3, {{0x00, 0xdef8}, {0x00, 0xdef8}, {0x00, 0xde
f8}, {0x00, 0xdef8}}, 0x00, 0},

seems to do it's job for bus1. bus2 has entries for device 3 and 4. Are
these correct for the current tree? 
All these infrastructural changes seem to cry for automatic irq-table
generation..

Another thing, when talking about tables: Is there anything serious 
speaking against including a minimal set of ACPI tables to get things
like HPET timers working? Or is there another viable way for enabling
these?

Stefan

[ron minnich]

On Tue, 4 Nov 2003, Stefan Reinauer wrote:
> Maybe an additional entry in irq_tables.c helps?
>  /* Let Linux know about bus 1 */
>  {0x01, (0x04<<3)|3, {{0x00, 0xdef8}, {0x00, 0xdef8}, {0x00, 0xde
> f8}, {0x00, 0xdef8}}, 0x00, 0},

that is already in there. It did not help. Linux is not scanning all the 
busses again.

Time for debug printks I guess.

ron