From mboxrd@z Thu Jan  1 00:00:00 1970
From: jdzheng@broadcom.com (Jiandong Zheng)
Date: Tue, 17 May 2011 16:42:26 -0700
Subject: [PATCH] ARM: bcmring: fix patches to convert to sp804 clockevents
In-Reply-To: <20110517232015.GC5913@n2100.arm.linux.org.uk>
References: <20110516172334.GD13659@n2100.arm.linux.org.uk>
	<E1QM1e8-0003ET-HK@rmk-PC.arm.linux.org.uk>
	<4DD2FF02.1020508@broadcom.com>
	<20110517232015.GC5913@n2100.arm.linux.org.uk>
Message-ID: <4DD307E2.7010302@broadcom.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On 5/17/2011 4:20 PM, Russell King - ARM Linux wrote:
> On Tue, May 17, 2011 at 04:04:34PM -0700, Jiandong Zheng wrote:
>> @@ -232,12 +230,14 @@ void __init bcmring_init_timer(void)
>>   	writel(0, TIMER2_VA_BASE + TIMER_CTRL);
>>   	writel(0, TIMER3_VA_BASE + TIMER_CTRL);
>>
>> +	clkdev_add_table(lookups, ARRAY_SIZE(lookups));
>> +
>
> As we have the init_early callback in the machine record, can this be
> done at that point, rather than sticking it in the timer init callback?
>
It should be OK as long as the table is added early enough but leave it 
here make it a little easier to understand the init order.

clkdev_add_table() was called by init_machine() which is too late while 
some other machines still do similar thing. Should we move 
clkdev_add_table() to init_early for all SoC using sp804?