[Lse-tech] RE: [PATCH] 2.6.5- es7000 subarch update

[Len B. <len...@in...>]

Natalie,

> ACPI: INT_SRC_OVR (bus 0 bus_irq 15 global_irq 13 high edge)

I agree with your description of why the 15->13 override fails.

> ACPI: INT_SRC_OVR (bus 0 bus_irq 4 global_irq 14 high edge)
> ACPI: INT_SRC_OVR (bus 0 bus_irq 14 global_irq 15 high edge)

These two also create an actual and potential duplicate mp_irqs[] entry.

> ACPI: INT_SRC_OVR (bus 0 bus_irq 6 global_irq 16 high edge)
> ACPI: INT_SRC_OVR (bus 0 bus_irq 7 global_irq 17 high edge)
> ACPI: INT_SRC_OVR (bus 0 bus_irq 8 global_irq 18 low edge)
> ACPI: INT_SRC_OVR (bus 0 bus_irq 12 global_irq 19 high edge)
> ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 20 high edge)
> ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 23 high level)

The ES7000 sure is a great tester of the over-ride code!

But I don't like the proposal to selectively invalidate
existing mp_irqs[] entries.

I think the proper fix is to parse the over-ride entries before
filling in the (remaining) identity mappings.  This also gets
rid of the special case for IRQ2, which would be handled exactly
like the mappings to < 16 on the ES7000 above.

Perhaps I should send you a patch you can test on the ES7000,
since I don't have one of those?

In any case, I'd prefer that proposed patches to this code come
through me, since it is ACPI specific.

thanks,
-Len