On 2016년 03월 28일 12:11, MyungJoo Ham wrote:
> []
>> Suggested-by: Myungjoo Ham <myungjoo.ham@xxxxxxxxxxx>
>> Signed-off-by: Chanwoo Choi <cw00.choi@xxxxxxxxxxx>
>> [tjakobi: Reported RCU locking issue and cw00.choi fix it.]
>> Reported-by: Tobias Jakobi <tjakobi@xxxxxxxxxxxxxxxxxxxxx>
>> [m.reichl and linux.amoon: Tested it on exynos4412-odroidu3 board]
>> Tested-by: Markus Reichl <m.reichl@xxxxxxxxxxxxx>
>> Tested-by: Anand Moon <linux.amoon@xxxxxxxxx>
>> ---
>> drivers/devfreq/Kconfig | 7 ++
>> drivers/devfreq/Makefile | 1 +
>> drivers/devfreq/devfreq.c | 17 ++++
>> drivers/devfreq/governor.h | 15 +++
>> drivers/devfreq/governor_passive.c | 192 +++++++++++++++++++++++++++++++++++++
>> include/linux/devfreq.h | 3 +
>> 6 files changed, 235 insertions(+)
>> create mode 100644 drivers/devfreq/governor_passive.c
>>
>> diff --git a/drivers/devfreq/Kconfig b/drivers/devfreq/Kconfig
>
> []
>
>> diff --git a/drivers/devfreq/Makefile b/drivers/devfreq/Makefile
>> index 8af8aaf922a8..2633087d5c63 100644
>> --- a/drivers/devfreq/Makefile
>> +++ b/drivers/devfreq/Makefile
>> @@ -4,6 +4,7 @@ obj-$(CONFIG_DEVFREQ_GOV_SIMPLE_ONDEMAND) += governor_simpleondemand.o
>> obj-$(CONFIG_DEVFREQ_GOV_PERFORMANCE) += governor_performance.o
>> obj-$(CONFIG_DEVFREQ_GOV_POWERSAVE) += governor_powersave.o
>> obj-$(CONFIG_DEVFREQ_GOV_USERSPACE) += governor_userspace.o
>> +obj-$(CONFIG_DEVFREQ_GOV_PASSIVE) += governor_passive.o
>>
>> # DEVFREQ Drivers
>> obj-$(CONFIG_ARM_EXYNOS_BUS_DEVFREQ) += exynos-bus.o
>> diff --git a/drivers/devfreq/devfreq.c b/drivers/devfreq/devfreq.c
>> index 1d6c803804d5..9f84bbc2994c 100644
>> --- a/drivers/devfreq/devfreq.c
>> +++ b/drivers/devfreq/devfreq.c
>> @@ -478,7 +478,13 @@ static void _remove_devfreq(struct devfreq *devfreq)
>> dev_warn(&devfreq->dev, "releasing devfreq which doesn't exist\n");
>> return;
>> }
>> +
>> + if (devfreq->governor->type == DEVFREQ_GOV_PASSIVE)
>> + devfreq_passive_unregister_notifier(devfreq);
>> +
>> list_del(&devfreq->node);
>> + list_del_init(&devfreq->passive_node);
>> +
>
> Let's do this inside governor_passive.c, you already have
> the DEVFREQ_GOV_STOP signal. This may be called more frequently
> because the START-STOP pair has wider usage than add-remove pair.
> However, still, it's ok to be detached during the "STOP"ed but not
> removed state.
I agree the code about passive governor to drivers/devfreq/governor_passive.c.
It makes the code clear.
>
>> mutex_unlock(&devfreq_list_lock);
>>
>> if (devfreq->governor)
>> @@ -598,6 +604,17 @@ struct devfreq *devfreq_add_device(struct device *dev,
>> goto err_init;
>> }
>>
>> + if (devfreq->governor->type == DEVFREQ_GOV_PASSIVE) {
>> + struct devfreq *parent_devfreq = (struct devfreq *)data;
>> + list_add(&devfreq->passive_node, &parent_devfreq->passive_node);
>> +
>> + err = devfreq_passive_register_notifier(devfreq);
>> + if (err <0)
>> + goto err_init;
>> + } else {
>> + INIT_LIST_HEAD(&devfreq->passive_node);
>> + }
>> +
>
> Same as above. We can reuse DEVFREQ_GOV_START for this.
> With DEVFREQ_GOV_START/STOP, we can entirely remove any modifications
> in the devfreq.c, governor.h, devfreq.h.
I agree. I'll modify it.
>
> Besides, such an approach removed the need for the patch 05/21.
> We no longer (or yet) need "governor type".
>
>> return devfreq;
>>
>> err_init:
>> diff --git a/drivers/devfreq/governor.h b/drivers/devfreq/governor.h
>> index cf19b923c362..64d1dffcdb43 100644
>> --- a/drivers/devfreq/governor.h
>> +++ b/drivers/devfreq/governor.h
>
> as mentioned above, we don't need to modify this file.
OK.
>
>> diff --git a/drivers/devfreq/governor_passive.c b/drivers/devfreq/governor_passive.c
>> new file mode 100644
>> index 000000000000..521e93b68c11
>> --- /dev/null
>> +++ b/drivers/devfreq/governor_passive.c
>> @@ -0,0 +1,192 @@
>> +/*
>> + * linux/drivers/devfreq/governor_passive.c
>> + *
>> + * Copyright (C) 2016 Samsung Electronics
>> + * Author: Chanwoo Choi <cw00.choi@xxxxxxxxxxx>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + */
>> +
>> +#include <linux/devfreq.h>
>> +#include <linux/module.h>
>> +#include <linux/device.h>
>> +#include <linux/devfreq.h>
>
> Doubly included
I'll fix it.
>
>> +#include "governor.h"
>> +
>
> []
>
>> +static int devfreq_passive_event_handler(struct devfreq *devfreq,
>> + unsigned int event, void *data)
>> +{
>> + return 0;
>
> Let's handle DEVFREQ_GOV_START/STOP event here.
OK. I'll register/unregister the DEVFREQ_TRANSITION_NOTIFIER by using
DEVFREQ_GOV_START/STOP event.
>
>> +}
>> +
>> +static struct devfreq_governor devfreq_passive = {
>> + .name = "passive",
>> + .type = DEVFREQ_GOV_PASSIVE,
>
> Let's not use .type enum, yet. We don't this this. At least for now.
OK. I'll remove it.
>
>> + .get_target_freq = devfreq_passive_get_target_freq,
>> + .event_handler = devfreq_passive_event_handler,
>> +};
>> +
>
> []
>
>> diff --git a/include/linux/devfreq.h b/include/linux/devfreq.h
When the devfreq device using passive governor registers the notifier
to DEVFREQ_TRANSITION_NOTIFIER, the devfreq device need the one notifier block field
as following:
struct devfreq {
...
...
struct notifier_block passive_nb;
}
static int devfreq_passive_event_handler(struct devfreq *devfreq,
unsigned int event, void *data)
{
struct device *dev = devfreq->dev.parent;
struct devfreq *parent = (struct devfreq *)devfreq->data;
struct notifier_block *nb = &devfreq->passive_nb;
int ret = 0;
if (!parent)
return -EPROBE_DEFER;
switch (event) {
case DEVFREQ_GOV_START:
nb->notifier_call = devfreq_passive_notifier_call;
ret = devm_devfreq_register_notifier(dev, parent, nb,
DEVFREQ_TRANSITION_NOTIFIER);
break;
case DEVFREQ_GOV_STOP:
devm_devfreq_unregister_notifier(dev, parent, nb,
DEVFREQ_TRANSITION_NOTIFIER);
break;
default:
break;
}
return ret;
}
Best Regards,
Chanwoo Choi
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
