* Tony Lindgren <tony@atomide.com> [081031 12:21]:
> In the case of spurious interrupt, the handler for previous irq most likely
> needs a read back of some register to ensure write posting.

Here's a better version to kill the checkpatch.pl warning. Later on we
might want to add also a check for "irq == 95" to the test in
omap_mask_irq() so we can keep this patch around with minimal overhead.

Tony