Has anyone tested the PIRQ table generated by the dbm690t code?
With a really trimmed down kernel, no ACPI and IO_APIC disabled the kernel goes looking for the PIRQ table. It finds one, but something isn't right the interrupts don't get routed. ( I really LOVE how there are three different ways to get the same information. )
Comparing the generated PIRQ to one pulled from my BIOS, they are nowhere near similar. I've tried replacing the dbm690t PIRQ table generator with a table from my BIOS, but it tells me the CRC is wrong. Something smells a little fishy.
The mptables are very similar as expected. By telling the kernel to use the advanced IO_APIC, I can finally boot to a prompt.
Next issue to resolve... X11 with fglrx... Is this the right list to post FILO patches?
I've attached my PIRQ and MPtables for your reading enjoyment.
Dan Lykowski
I think the lesson is "avoid PIRQ tables". Let's just get the MP table right.
I'm going to start work on this next week as I hope to have a nice demo for SCALE. So your work is important to me :-)
ron
On Sat, Jan 24, 2009 at 10:23 AM, ron minnich rminnich@gmail.com wrote:
I think the lesson is "avoid PIRQ tables". Let's just get the MP table right.
I'm going to start work on this next week as I hope to have a nice demo for SCALE. So your work is important to me :-)
i don't think that the PIRQ table was tested. The MP table should be correct and is the one expected to be used. It would be nice if all the tables could be generated from the same source table within coreboot.
Marc
On Sat, Jan 24, 2009 at 9:34 AM, Marc Jones marcj303@gmail.com wrote:
i don't think that the PIRQ table was tested. The MP table should be correct and is the one expected to be used. It would be nice if all the tables could be generated from the same source table within coreboot.
I agree but in today's world, where PIRQ handlers in operating systems don't always do the right thing, I would rather we move as quickly as possible to stop using/supporting PIRQ tables at all (where that is possible; I realize that on some systems PIRQ is mandatory, but certainly not on K8 platforms).
Thanks
ron
ron minnich wrote:
On Sat, Jan 24, 2009 at 9:34 AM, Marc Jones marcj303@gmail.com wrote:
i don't think that the PIRQ table was tested. The MP table should be correct and is the one expected to be used. It would be nice if all the tables could be generated from the same source table within coreboot.
I agree but in today's world, where PIRQ handlers in operating systems don't always do the right thing, I would rather we move as quickly as possible to stop using/supporting PIRQ tables at all (where that is possible; I realize that on some systems PIRQ is mandatory, but certainly not on K8 platforms).
PIRQ is only required when we don't have ACPI. Once we have ACPI on all platforms we can drop PIRQ.
-
Dan Lykowski -
Marc Jones -
ron minnich -
Stefan Reinauer