On Thu, Mar 1, 2012 at 5:51 PM, Arnd Bergmann <[email protected]> wrote: > On Thursday 01 March 2012, Haojian Zhuang wrote: >> +#ifdef CONFIG_OF >> +static int sa1100_rtc_probe_dt(struct platform_device *pdev, >> + struct sa1100_rtc *info) >> +{ >> + info->irq_alarm = platform_get_irq(pdev, 0); >> + info->irq_1hz = platform_get_irq(pdev, 1); >> + if (info->irq_1hz < 0 || info->irq_alarm < 0) >> + return -ENODEV; >> + return 0; >> +} >> +#else >> +static int sa1100_rtc_probe_dt(struct platform_device *pdev, >> + struct sa1100_rtc *info) >> +{ >> + return 1; >> +} >> +#endif >> + >> +static int sa1100_rtc_probe_irq(struct platform_device *pdev, >> + struct sa1100_rtc *info) >> +{ >> + info->irq_1hz = platform_get_irq_byname(pdev, "rtc 1Hz"); >> + info->irq_alarm = platform_get_irq_byname(pdev, "rtc alarm"); >> + if (info->irq_1hz < 0 || info->irq_alarm < 0) >> + return -ENODEV; >> + return 0; >> +} > > This should not be necessary. I thought we had already added support > for named irq resources coming from the device tree. If not, why not > just make the other ones anonymous as well and just use platform_get_irq()? >
Yes, platform_get_irq() is better. I'll append the two irqs into interrupts property. _______________________________________________ devicetree-discuss mailing list [email protected] https://lists.ozlabs.org/listinfo/devicetree-discuss
