[rtc-linux] [PATCH] rtc/ds3232: Enable ds3232 to work as wakeup source

[Dongsheng.Wang at freescale.com]

> -----Original Message-----
> From: Wood Scott-B07421
> Sent: Wednesday, February 26, 2014 11:21 AM
> To: Wang Dongsheng-B40534
> Cc: Andrew Morton; rtc-linux at googlegroups.com; benh at kernel.crashing.org;
> a.zummo at towertech.it; Zhao Chenhui-B35336; linuxppc-dev at lists.ozlabs.org
> Subject: Re: [rtc-linux] [PATCH] rtc/ds3232: Enable ds3232 to work as wakeup
> source
> 
> On Tue, 2014-02-25 at 21:09 -0600, Wang Dongsheng-B40534 wrote:
> >
> > > -----Original Message-----
> > > From: Andrew Morton [mailto:akpm at linux-foundation.org]
> > > Sent: Wednesday, February 26, 2014 6:07 AM
> > > To: rtc-linux at googlegroups.com
> > > Cc: Wang Dongsheng-B40534; a.zummo at towertech.it; Zhao Chenhui-B35336;
> linuxppc-
> > > dev at lists.ozlabs.org
> > > Subject: Re: [rtc-linux] [PATCH] rtc/ds3232: Enable ds3232 to work as wakeup
> > > source
> > >
> > > On Tue, 21 Jan 2014 13:24:51 +0800 Dongsheng Wang
> <dongsheng.wang at freescale.com>
> > > wrote:
> > >
> > > > +	if (client->irq != NO_IRQ) {
> > >
> > > x86_64 allmodconfig:
> > >
> > > drivers/rtc/rtc-ds3232.c: In function 'ds3232_probe':
> > > drivers/rtc/rtc-ds3232.c:427: error: 'NO_IRQ' undeclared (first use in this
> > > function)
> > > drivers/rtc/rtc-ds3232.c:427: error: (Each undeclared identifier is reported
> > > only once
> > > drivers/rtc/rtc-ds3232.c:427: error: for each function it appears in.)
> > >
> > > Not all architectures implement NO_IRQ.
> > >
> > > I think this should be
> > >
> > > 	if (client->irq > 0) {
> > >
> > > but I'm not sure - iirc, x86 (at least) treats zero as "not an IRQ".
> > > But I think some architectures permit IRQ 0.  There was discussion many
> > > years ago but I don't think anything got resolved.
> > >
> > I think this is why NO_IRQ is defined in kernel, that should be resolved this
> issue.
> >
> > Sorry, I don't know why some architectures didn't define this macro?
> 
> NO_IRQ is deprecated (see "git log -SNO_IRQ" for the trend of removing
> uses of it, as well as situations where it gives the wrong results).
> "if (client->irq > 0)" is correct.
> 
Thanks.

-Dongsheng

> -Scott
>