[RFC PATCHv2 14/20] ARM: omap4: dynamically register local timersetup function

[Marc Zyngier]

Hi Santosh,

On Wed, 2011-03-16 at 11:25 +0530, Santosh Shilimkar wrote:
> Marc,
> > -----Original Message-----
> > From: linux-arm-kernel-bounces at lists.infradead.org [mailto:linux-
> > arm-kernel-bounces at lists.infradead.org] On Behalf Of Marc Zyngier
> > Sent: Tuesday, March 15, 2011 8:43 PM
> > To: linux-arm-kernel at lists.infradead.org
> > Cc: Tony Lindgren
> > Subject: [RFC PATCHv2 14/20] ARM: omap4: dynamically register local
> > timersetup function
> >
> > Cc: Tony Lindgren <tony at atomide.com>
> > Signed-off-by: Marc Zyngier <marc.zyngier at arm.com>
> > ---
> >  arch/arm/mach-omap2/Makefile    |    1 -
> >  arch/arm/mach-omap2/timer-gp.c  |   13 ++++++++++++-
> >  arch/arm/mach-omap2/timer-mpu.c |   36 ----------------------------
> > --------
> >  3 files changed, 12 insertions(+), 38 deletions(-)
> >  delete mode 100644 arch/arm/mach-omap2/timer-mpu.c
> >
> > diff --git a/arch/arm/mach-omap2/Makefile b/arch/arm/mach-
> > omap2/Makefile
> > index a45cd64..41f8b67 100644
> > --- a/arch/arm/mach-omap2/Makefile
> > +++ b/arch/arm/mach-omap2/Makefile
> > @@ -22,7 +22,6 @@ obj-$(CONFIG_TWL4030_CORE) += omap_twl.o
> >
> >  # SMP support ONLY available for OMAP4
> >  obj-$(CONFIG_SMP)			+= omap-smp.o omap-headsmp.o
> > -obj-$(CONFIG_LOCAL_TIMERS)		+= timer-mpu.o
> >  obj-$(CONFIG_HOTPLUG_CPU)		+= omap-hotplug.o
> >  obj-$(CONFIG_ARCH_OMAP4)		+= omap44xx-smc.o omap4-common.o
> >
> > diff --git a/arch/arm/mach-omap2/timer-gp.c b/arch/arm/mach-
> > omap2/timer-gp.c
> > index 3b9cf85..98d03c5 100644
> > --- a/arch/arm/mach-omap2/timer-gp.c
> > +++ b/arch/arm/mach-omap2/timer-gp.c
> > @@ -39,6 +39,7 @@
> >  #include <asm/mach/time.h>
> >  #include <plat/dmtimer.h>
> >  #include <asm/localtimer.h>
> > +#include <asm/smp_twd.h>
> >  #include <asm/sched_clock.h>
> >  #include <plat/common.h>
> >  #include <plat/omap_hwmod.h>
> > @@ -247,12 +248,22 @@ static void __init
> > omap2_gp_clocksource_init(void)
> >  }
> >  #endif
> >
> > +#ifdef CONFIG_LOCAL_TIMERS
> > +static int __cpuinit omap4_local_timer_setup(struct
> > clock_event_device *evt)
> > +{
> > +	evt->irq = OMAP44XX_IRQ_LOCALTIMER;
> > +	return 0;
> > +}
> > +#endif
> > +
> >  static void __init omap2_gp_timer_init(void)
> >  {
> >  #ifdef CONFIG_LOCAL_TIMERS
> > -	if (cpu_is_omap44xx()) {
> > +	if (cpu_is_omap44xx() && omap_rev() != OMAP4430_REV_ES1_0)
> 
> I see one problem here. The reason we changed the signature of
> local_timer_setup() is in case where the local timer isn't
> available on a specific SOC versions, the kernel will switch
> to broad-cast timers.
> 
> This way the kernel built with CONFIG_LOCAL_TIMERS can still work
> on SoCs where twd isn't supported.
> 
> Looks like your patch breaks that unless and until I am missing
> something.

I think the bit you're missing is that when no local timer is
registered, the code in percpu_timer.c::percpu_timer_setup() ensures
that the broadcast timers are used as a default.

In the OMAP4 case, twd is registered unless we detect a
OMAP4430_REV_ES1_0. In that case, we let the magic happen and fall back
to the broadcast timers. Tested on a PandaBoard with twd artificially
disabled.

	M.
-- 
Reality is an implementation detail.