On Sun, Feb 13, 2011 at 02:00:26PM -0800, David Miller wrote:
> From: Sam Ravnborg <sam@xxxxxxxxxxxx>
> Date: Sun, 13 Feb 2011 22:48:44 +0100
>
> > In the original implmentation we did the same via __enable_irq().
>
> And we did so Via a function that's invoked by every cpu as it is booted.
>
> Right?
To the best of my knowledge - no.
In irq_32.c:request_irq() we had:
__enable_irq(irq); // irq is what we consider real_irq
In irq.h: we had:
BTFIXUPDEF_CALL(void, enable_irq, unsigned int)
static inline void __enable_irq(unsigned int irq)
{
BTFIXUP_CALL(enable_irq)(irq);
}
In sun4m_irq.c we had:
void __init sun4m_init_IRQ(void)
{
...
BTFIXUPSET_CALL(enable_irq, sun4m_enable_irq, BTFIXUPCALL_NORM);
So unless the btfix stuff does trick I do not understand then I
do not see this iterate over all cpu's.
I guess there is something obvious I have missed :-(
One thing I have left-out on the sun4m side is a call
to smp4m_irq_rotate() that for each interrupt let
the next be the interrupt target.
But I do not see that interrupts are globally enabled.
Note: I got one step further. It helped that I
called irq_ack(), irq_enable() around my desc->handle_irq(irq, desc);
But desc->handle_irq in NULL.
That will be for another day - it is getting late.
Thanks for all the help so far!
Sam
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
