From mboxrd@z Thu Jan  1 00:00:00 1970
From: joelf@ti.com (Joel Fernandes)
Date: Fri, 22 Nov 2013 10:01:41 -0600
Subject: [PATCH 1/8] ARM: OMAP: Move public portion of dmtimer.h to
 include/linux/omap-timer.h
In-Reply-To: <20131122153334.GE10023@atomide.com>
References: <1385085414-9034-1-git-send-email-joelf@ti.com>
 <1385085414-9034-2-git-send-email-joelf@ti.com>
 <20131122153334.GE10023@atomide.com>
Message-ID: <528F7FE5.3020908@ti.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On 11/22/2013 09:33 AM, Tony Lindgren wrote:
> Hi,
> 
> * Joel Fernandes <joelf@ti.com> [131121 18:00]:
>> Multiplatform support has made arch/arm/plat-omap/include/plat/ inaccessible to
>> drivers outside the plat-omap directory [1]. Due to this the following drivers
>> are disabled with !CONFIG_ARCH_MULTIPLATFORM:
>> CONFIG_IR_RX51 (drivers/media/rc/ir-rx51.c)
>> CONFIG_TIDSPBRIDGE (drivers/staging/tidspbridge/core/dsp-clock.c)
>>
>> We move the portion of the dmtimer "API" that should be accessible to the
>> drivers, into include/linux/omap-timer.h
> 
> As we chatted earlier, we don't have to expose all these hardware specific
> functions and use existing Linux generic frameworks instead.
> 
> We can implement an irqchip and a clocksource in the dmtimer code for the
> client drivers to use, and after that we only have a couple of dmtimer
> specific functions left to export.
> 
> I'm thinkging some thing like this for the public API:
> 
> omap_dm_timer_request			request_irq
> omap_dm_timer_request_specific		request_irq
> omap_dm_timer_get_irq			request_irq
> omap_dm_timer_set_source		clk_set_rate
> omap_dm_timer_stop			disable_irq
> omap_dm_timer_start			enable_irq
> omap_dm_timer_disable			disable_irq
> omap_dm_timer_free			free_irq
> omap_dm_timer_set_int_enable		enable_irq

Hi Tony,

The thing I feel we're trying to fit a square peg into round hole type of thing
here.

Some reasons I feel this is overkill (point 2 makes it even not possible):

1. Ideally IR_RX51 should be using clockevents/hrtimer framework instead of
irqchip. Other than that, dsp bridge is the only other user.

2. These functions will be also be required to be public once mach-omap2/timer.c
moves to drivers/clocksource/ . At such an early point of code, we can't use
irqchip anyway so we'd need these functions public.

3. This is a minor point, but its also not immediately intuitive or clear to
someone who needs a dedicated hardware timer that they need to use an IRQ chip
driver for the same. But we can avoid any discussion about this till we can
discuss/agree on the above 2.

thanks,

-Joel