On Tue, Aug 10, 2010 at 02:00:55PM +0200, Uwe Kleine-König wrote: > Hi Wolfram, > > > > + retval = request_irq(rtc->irq, lpc32xx_rtc_alarm_interrupt, > > > + IRQF_DISABLED, "rtcalarm", rtc); > > > + if (retval < 0) { > > > + dev_err(&pdev->dev, "Can't request interrupt\n"); > > > + goto err_free_irq; > > > + } > > > > I saw that a number of rtc-drivers register their irq after they > > register the device. I wonder if this is OK here? Couldn't it happen > > that after rtc_device_register() there is a preemption and another > > process could set the alarm? Then there is a race between interrupts > > already enabled and no handler available, no? > If you do it the other way around the irq might trigger and the handler > reports an irq for a device that doesn't exist yet. Well, I was assuming that you initially have all interrupts disabled... -- Pengutronix e.K. | Wolfram Sang | Industrial Linux Solutions | http://www.pengutronix.de/ |