Hi,

On Wed, Jun 29, 2011 at 04:01:22PM -0400, Joey Oravec wrote:

Nicolas, Lennert -

I just tested linux kernel 3.0-rc5 today. There were some major
changes to arch/arm/plat-orion/gpio.c and I think they introduced a
bug for edge sensitive interrupts. I have not tested any other types
yet. This new version of the code specifies

As level interrupts don't need to be acked, it should be ok.

ct->chip.irq_ack = irq_gc_ack;

But it looks like kernel/irq/generic-chip.c generates a positive
mask (1 to clear) and the MV78200 wants the opposite (0 to clear).
The result is that the interrupt never gets ACKed and the handler
will keep getting called.

Maybe we could introduce some irq_gc_ack{_set_bit,_clear_bit}
variants (as for the mask functions) ?

As an another option, we could define an Orion specific irq_ack
function.

I don't know what is the better choice for rc5.

Regards,

Simon
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20110630/b77176e8/attachment.sig>