Subject: Re: CVS commit: src/sys/arch/arm/xscale
To: Steve Woodford <scw@netbsd.org>
From: David Laight <david@l8s.co.uk>
List: source-changes
Date: 10/08/2003 22:04:10
> Log Message:
> Ok, I give up for now. There's no easy/reliable way to deal with
> these spurious interrupts.

Are they caused by the PCI write that should clear the IRQ being
buffered (eg a posted write on a PCI bridge), and not being executed
on the IO card until after the CPU has returned from the interrupt?

If so, then they are fixable by doing a read-back of the PCI write.

	David

-- 
David Laight: david@l8s.co.uk