On 30 September 2014 12:36, Arnd Bergmann <arnd@xxxxxxxx> wrote: > On Tuesday 30 September 2014 12:22:26 Rafał Miłecki wrote: >> @@ -218,6 +218,12 @@ int bcma_gpio_init(struct bcma_drv_cc *cc) >> #if IS_BUILTIN(CONFIG_BCM47XX) >> chip->to_irq = bcma_gpio_to_irq; >> #endif >> +#if IS_BUILTIN(CONFIG_OF) >> + if (cc->core->bus->hosttype == BCMA_HOSTTYPE_SOC) >> + chip->of_node = of_find_compatible_node( >> + bus->host_pdev->dev.of_node, NULL, >> + "brcm,bus-chipcommon"); >> +#endif >> switch (cc->core->bus->chipinfo.id) { > > This doesn't: you are now searching through all nodes starting at the > axi node rather than searching just through the children. > > I think it would be better with the first change in place to set > chip->of_node to cc->core->dev.of_node, and set that pointer in > bcma_bus_scan by matching the 'reg' number. I think that is what > an earlier version of the bcma DT support did in order to find the > IRQs. We no longer need it for that purpose, but it seems like a > good idea anyway, as I expect other bcma_devices to have similar > requirements to add additional properties. Ohh, of course, I didn't notice that [PATCH v6] bcma: register bcma as device tree driver already contains bcma_of_fill_device. This will simplify my search a lot! -- Rafał -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html