Re: Oops in UHCI when encountering "host controller process error"

[Jeremy Fitzhardinge <jeremy@goop.org>]

Alan Stern wrote:
> The entry point is uhci_urb_enqueue().  You can do whatever you like 
> there, but bear in mind that it may be called in interrupt context.  
> uhci_sprint_schedule() should be fine.
>
> However I don't really see the point.  If the DMA pool allocations are 
> so badly messed up that the initial setup is wrong, why bother to 
> search any further?

I was trying to use it as a way of gauging where the damage was coming from.

But its moot now; Ian Campbell tracked it down to a missing TLB flush 
which explained all the weird non-deterministic symptoms we were 
seeing.  It all looks pretty good now.

    J