Re: [PATCH v1] i3c: master: Enable runtime PM for master controller

[Mukesh Kumar Savaliya <quic_msavaliy@quicinc.com>]

Hi Joshua,

On 3/15/2024 1:35 PM, Joshua Yeong wrote:
> 
> 
>> -----Original Message-----
>> From: linux-i3c <linux-i3c-bounces@lists.infradead.org> On Behalf Of Mukesh
>> Kumar Savaliya
>> Sent: Monday, March 11, 2024 1:30 AM
>> To: alexandre.belloni@bootlin.com; linux-i3c@lists.infradead.org; linux-
>> kernel@vger.kernel.org
>> Cc: andersson@kernel.org; vkoul@kernel.org;
>> manivannan.sadhasivam@linaro.org
>> Subject: Re: [PATCH v1] i3c: master: Enable runtime PM for master controller
>>
>> A Gentle reminder ! As i was not part of the linux-i3c group before raising this
>> gerrit, might not went into right folder.
>>
>> On 2/28/2024 3:04 PM, Mukesh Kumar Savaliya wrote:
>>> Enable runtime PM for i3c master node during master registration time.
>>>
>>> Sometimes i3c client device driver may want to control the PM of the
>>> parent (master) to perform the transactions and save the power in an
>>> efficient way by controlling the session. Hence device can call PM
>>> APIs by passing the parent node.
>>>
>>> Here, I3C target device when calls pm_runtime_get_sync(dev->parent)
>>> couldn't invoke master drivers runtime PM callback registered by the
>>> master driver because parent's PM status was disabled in the Master
>>> node.
>>>
>>> Also call pm_runtime_no_callbacks() and pm_suspend_ignore_children()
>>> for the master node to not have any callback addition and ignore the
>>> children to have runtime PM work just locally in the driver. This
>>> should be generic and common change for all i3c devices and should not
>>> have any other impact.
>>>
>>> With these changes, I3C client device works and able to invoke master
>>> driver registered runtime PM callbacks.
>>>
>>> Signed-off-by: Mukesh Kumar Savaliya <quic_msavaliy@quicinc.com>
>>> ---
>>>    drivers/i3c/master.c | 6 ++++++
>>>    1 file changed, 6 insertions(+)
>>>
>>> diff --git a/drivers/i3c/master.c b/drivers/i3c/master.c index
>>> 3afa530c5e32..a3dc88974f92 100644
>>> --- a/drivers/i3c/master.c
>>> +++ b/drivers/i3c/master.c
>>> @@ -13,6 +13,7 @@
>>>    #include <linux/kernel.h>
>>>    #include <linux/list.h>
>>>    #include <linux/of.h>
>>> +#include <linux/pm_runtime.h>
>>>    #include <linux/slab.h>
>>>    #include <linux/spinlock.h>
>>>    #include <linux/workqueue.h>
>>> @@ -2812,6 +2813,10 @@ int i3c_master_register(struct
>>> i3c_master_controller *master,
>>>
>>>    	i3c_bus_notify(i3cbus, I3C_NOTIFY_BUS_ADD);
>>>
>>> +	pm_runtime_no_callbacks(&master->dev);
>>> +	pm_suspend_ignore_children(&master->dev, true);
>>> +	pm_runtime_enable(&master->dev);
>>> +
> 
> Will runtime pm impact on ibi request from target?
Don't think so. This change just enables target device driver to call PM 
apis and control the transfers under RT resume and RT suspend.
IBI comes asynchronously to the master driver and it should get handled 
accordingly. .
> 
>>>    	/*
>>>    	 * We're done initializing the bus and the controller, we can now
>>>    	 * register I3C devices discovered during the initial DAA.
>>> @@ -2849,6 +2854,7 @@ void i3c_master_unregister(struct
>> i3c_master_controller *master)
>>>    	i3c_master_i2c_adapter_cleanup(master);
>>>    	i3c_master_unregister_i3c_devs(master);
>>>    	i3c_master_bus_cleanup(master);
>>> +	pm_runtime_disable(&master->dev);
>>>    	device_unregister(&master->dev);
>>>    }
>>>    EXPORT_SYMBOL_GPL(i3c_master_unregister);
>>
>> --
>> linux-i3c mailing list
>> linux-i3c@lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/linux-i3c