From mboxrd@z Thu Jan  1 00:00:00 1970
From: acourbot@nvidia.com (Alexandre Courbot)
Date: Wed, 12 Feb 2014 10:56:18 +0900
Subject: [PATCH v2 6/6] ARM: tegra: cpuidle: use firmware for power down
In-Reply-To: <20140212013650.GA7584@quad.lixom.net>
References: <1391747706-1847-1-git-send-email-acourbot@nvidia.com>
 <1391747706-1847-7-git-send-email-acourbot@nvidia.com>
 <20140212013650.GA7584@quad.lixom.net>
Message-ID: <52FAD4C2.80803@nvidia.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On 02/12/2014 10:36 AM, Olof Johansson wrote:
> On Fri, Feb 07, 2014 at 01:35:06PM +0900, Alexandre Courbot wrote:
>> Attempt to invoke the prepare_idle() and do_idle() firmware calls
>> to power down a CPU so an underlying firmware gets informed of
>> the idle operation and performs it by itself if designed in such a way.
>>
>> Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
>> ---
>>   arch/arm/mach-tegra/cpuidle-tegra114.c | 11 ++++++++++-
>>   1 file changed, 10 insertions(+), 1 deletion(-)
>>
>> diff --git a/arch/arm/mach-tegra/cpuidle-tegra114.c b/arch/arm/mach-tegra/cpuidle-tegra114.c
>> index e0b87300243d..558067ddc186 100644
>> --- a/arch/arm/mach-tegra/cpuidle-tegra114.c
>> +++ b/arch/arm/mach-tegra/cpuidle-tegra114.c
>> @@ -19,6 +19,7 @@
>>   #include <linux/cpuidle.h>
>>   #include <linux/cpu_pm.h>
>>   #include <linux/clockchips.h>
>> +#include <asm/firmware.h>
>>
>>   #include <asm/cpuidle.h>
>>   #include <asm/suspend.h>
>> @@ -45,7 +46,15 @@ static int tegra114_idle_power_down(struct cpuidle_device *dev,
>>
>>   	clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ENTER, &dev->cpu);
>>
>> -	cpu_suspend(0, tegra30_sleep_cpu_secondary_finish);
>> +	call_firmware_op(prepare_idle);
>> +
>> +	switch (call_firmware_op(do_idle)) {
>> +	case -ENOSYS:
>> +		cpu_suspend(0, tegra30_sleep_cpu_secondary_finish);
>> +		break;
>> +	default:
>> +		break;
>
> Do you expect other cases down the road? If not, this is a simple if instead:
>
> 	/* Only call cpu_suspend if TF didn't handle the pre-suspend logic */
> 	if (call_firmware_op(do_idle) == -ENOSYS)
> 		cpu_suspend(....);

I might have been overdoing it indeed. Let me submit a v3 just for this 
one patch.

Thanks!
Alex.