On Fri, Dec 16, 2011 at 2:03 AM, Grazvydas Ignotas <notasas@xxxxxxxxx> wrote: > I'm running 3.2-rc5 on pandora with Neil's twl patches on top, and > omap_hsmmc driver fails to probe: > [ 3.491394] omap_hsmmc omap_hsmmc.0: context was not lost > [ 3.497161] omap_hsmmc omap_hsmmc.0: enabled > [ 3.715270] omap_hsmmc omap_hsmmc.0: Unable to grab MMC CD IRQ > > Pandora is using twl4030 GPIOs for card detect. From what I tried to > debug, omap_hsmmc is failing after call to request_irq() on card > detect irq with -EINVAL, irq number is 384 which is in range of other > twl4030 irqs. It looks like request_irq() is calling > request_threaded_irq() on ARM with NULL thread_fn, which in turn calls > __setup_irq() that calls irq_settings_is_nested_thread() that results > with true. As this nested flag is set it checks thread_fn, but that is > NULL causing -EINVAL final result. > > Could someone who knows irq code better look at this? I think this > should reproduce on beagle too. Basically all the modules of twl4030 should be converted to use request_threaded_irq(), if they are not, they will fail, and that's good, as trying to workaround the issue would just cause more trouble. -- Felipe Contreras -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html