> Okay, and here is my latest attempt to deal with the issue (which is
> dirty as sin code-wise, but nm that... I'm trying to debug this).
>
> This doesn't actually work, I end up deadlocking presumably because
> everything's waiting for an interrupt that's been dropped at some point.
> +int vector_to_irq[MAX_NUMNODES][FIRST_SYSTEM_VECTOR - FIRST_DEVICE_VECTOR + 1];
This can never work by looking at the I/O interrupt section of your MP
table you're overwriting interrupt gates. Therefore this has the
unfortunate effect of invalidating the rest of the patch.
Consider this; If you have irq 8 on both ioapic0 and ioapic3 who gets the
vector setup in the IDT?
Zwane
-- function.linuxpower.ca
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/