[V3 0/6] rtc-sa1100 clean up

[Russell King - ARM Linux]

On Thu, Jan 19, 2012 at 11:46:27AM +0000, Russell King - ARM Linux wrote:
> On Thu, Jan 19, 2012 at 11:39:40AM +0000, Russell King - ARM Linux wrote:
> > On Wed, Nov 30, 2011 at 02:34:03PM +0800, Jett.Zhou wrote:
> > > This patch set do the following things:
> > >   1) clean up sa1100-rtc for support sa1100/pxa/mmp series
> > >   2) add dummy clock for pxa arch
> > >   3) clean up clock dev for sa1100 arch and add dummy clock for sa1100
> > >   4) fix build error for sa1100 arch
> > 
> > Guys,
> > 
> > If you're modifying a driver which is shared across different architectures
> > please try to build it for those different architectures.  It helps to
> > avoid receiving reports like the below:
> > 
> > drivers/rtc/rtc-sa1100.c:42:1: warning: "RCNR" redefined
> > In file included from arch/arm/mach-sa1100/include/mach/hardware.h:73,
> >                  from drivers/rtc/rtc-sa1100.c:35:
> > arch/arm/mach-sa1100/include/mach/SA-1100.h:877:1: warning: this is the location of the previous definition
> > drivers/rtc/rtc-sa1100.c:43:1: warning: "RTAR" redefined
> > arch/arm/mach-sa1100/include/mach/SA-1100.h:876:1: warning: this is the location of the previous definition
> > drivers/rtc/rtc-sa1100.c:44:1: warning: "RTSR" redefined
> > arch/arm/mach-sa1100/include/mach/SA-1100.h:879:1: warning: this is the location of the previous definition
> > drivers/rtc/rtc-sa1100.c:45:1: warning: "RTTR" redefined
> > arch/arm/mach-sa1100/include/mach/SA-1100.h:878:1: warning: this is the location of the previous definition
> > drivers/rtc/rtc-sa1100.c:47:1: warning: "RTSR_HZE" redefined
> > arch/arm/mach-sa1100/include/mach/SA-1100.h:891:1: warning: this is the location of the previous definition
> > drivers/rtc/rtc-sa1100.c:48:1: warning: "RTSR_ALE" redefined
> > arch/arm/mach-sa1100/include/mach/SA-1100.h:890:1: warning: this is the location of the previous definition
> > drivers/rtc/rtc-sa1100.c:49:1: warning: "RTSR_HZ" redefined
> > arch/arm/mach-sa1100/include/mach/SA-1100.h:889:1: warning: this is the location of the previous definition
> > drivers/rtc/rtc-sa1100.c:50:1: warning: "RTSR_AL" redefined
> > arch/arm/mach-sa1100/include/mach/SA-1100.h:888:1: warning: this is the location of the previous definition
> > 
> > Please fix it.
> 
> And what's more, it doesn't even work on sa1100 anymore.  So it's regressed
> during this merge window.
> 
> Ok, this is a total bollocks change - what was added to this file
> was:
> 
> #define RCNR            0x00    /* RTC Count Register */
> #define RTAR            0x04    /* RTC Alarm Register */
> #define RTSR            0x08    /* RTC Status Register */
> #define RTTR            0x0c    /* RTC Timer Trim Register */
> 
> and the SA11x0 definitions are different:
> 
> #define RTAR            __REG(0x90010000)  /* RTC Alarm Reg. */
> #define RCNR            __REG(0x90010004)  /* RTC CouNt Reg. */
> #define RTTR            __REG(0x90010008)  /* RTC Trim Reg. */
> #define RTSR            __REG(0x90010010)  /* RTC Status Reg. */
> 
> Notice the different register positioning in the address space.
> 
> Therefore, this driver restructuring is fatally wrong.  I'm going to
> revert this entire series in my tree so SA11x0 continues to work here,
> whether I send that to Linus depends on how quickly a fix can be
> produced, and whether Alessandro agrees.

Moreover, the previous cleanup patch ("RTC: sa1100: remove redundant
code of setting alarm") wasn't tested without the other patches:

drivers/rtc/rtc-sa1100.c: In function 'sa1100_rtc_set_alarm':
drivers/rtc/rtc-sa1100.c:208: error: 'now' undeclared (first use in this function)
drivers/rtc/rtc-sa1100.c:208: error: (Each undeclared identifier is reported only once
drivers/rtc/rtc-sa1100.c:208: error: for each function it appears in.)
drivers/rtc/rtc-sa1100.c:210: error: incompatible type for argument 3 of 'rtc_next_alarm_time'
drivers/rtc/rtc-sa1100.c:211: error: 'time' undeclared (first use in this function)

so I've had to revert what should've been an independent change as well.

All these reverts are now sitting in my 'fixes' branch, and they are
the minimal set to restore the driver to a working state.