IRQ-Tables once again

Stefan Reinauer stepan at suse.de
Tue Jan 20 12:50:00 CET 2004


* ron minnich <rminnich at lanl.gov> [040120 16:59]:
> looks fine. 
> 
> Everyone is always having trouble. It may be best if we spend future
> effort on generating IRQ tables from MP tables, if that is possible.

I am not sure if our current mp tables know enough of the hardware to
do so. Interrupts are often hardwired so there has to be knowledge in
LinuxBIOS on which slot gets which interrupt. 

The configuration files should contain this knowledge. 

  < Bridges > -- 1:n -- < slots > -- 1:n(m:n?) -- < irqs >

Reading the Linux kernel sources it seems that Linux _does_ expect the
devices to have interrupts assigned already, which obviously LinuxBIOS
does not do. This has to be "fixed" if LinuxBIOS shall be open to
anybody not patching special kernels that do the right thing.

Is there any way to generate a decent mptable? The result from the
mptable utility in the freebios 1 tree seems not to do it's best on
AMD systems.

If we provide tables at all, can we maybe drop the irq table and only
use the mptable?

Best regards,
  Stefan Reinauer
  
-- 
     Stefan Reinauer, SUSE LINUX AG
Teamleader Architecture Development



More information about the coreboot mailing list