Re: [PATCH] thermal/drivers/tsens: Add suspend to RAM support for tsens

From: Priyansh Jain <quic_priyjain@quicinc.com>
To: Manaf Meethalavalappu Pallikunhi <quic_manafm@quicinc.com>,
	Konrad Dybcio <konrad.dybcio@linaro.org>,
	Amit Kucheria <amitk@kernel.org>,
	Thara Gopinath <thara.gopinath@gmail.com>,
	Bjorn Andersson <andersson@kernel.org>,
	"Rafael J . Wysocki" <rafael@kernel.org>,
	Daniel Lezcano <daniel.lezcano@linaro.org>,
	Zhang Rui <rui.zhang@intel.com>,
	Lukasz Luba <lukasz.luba@arm.com>, <linux-pm@vger.kernel.org>,
	<linux-arm-msm@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
	<quic_mkshah@quicinc.com>,
	"Raghavendra Kakarla" <quic_rkakarla@quicinc.com>
Subject: Re: [PATCH] thermal/drivers/tsens: Add suspend to RAM support for tsens
Date: Fri, 23 Feb 2024 10:50:34 +0530	[thread overview]
Message-ID: <ef3f5a8a-f5c7-3a2c-191b-4ce0db5cd79f@quicinc.com> (raw)
In-Reply-To: <1cd754e5-fc5e-bd8b-1d70-8de40c9a85e7@quicinc.com>

On 1/27/2024 9:07 PM, Manaf Meethalavalappu Pallikunhi wrote:

> +Maulik and Raghavendra
> 
> Hi Konrad,
> 
> On 1/25/2024 4:38 PM, Konrad Dybcio wrote:
>>
>>
>> On 1/24/24 16:25, Priyansh Jain wrote:
>>>
>>>
>>> On 1/24/2024 6:04 PM, Konrad Dybcio wrote:
>>>>
>>>>
>>>> On 1/24/24 11:42, Priyansh Jain wrote:
>>>>>
>>>>>
>>>>> On 1/22/2024 8:02 PM, Konrad Dybcio wrote:
>>>>>> On 22.01.2024 11:07, Priyansh Jain wrote:
>>>>>>> Add suspend callback support for tsens which disables tsens 
>>>>>>> interrupts
>>>>>>> in suspend to RAM callback.
>>>>>>
>>>>>> Would it not be preferrable to have the "critical overheat", wakeup-
>>>>>> capable interrupts be enabled, even if the system is suspended?
>>>>>>
>>>>>
>>>>>
>>>>> As part of suspend to RAM, tsens hardware will be turned off and it 
>>>>> cannot generate any interrupt.Also system doesn't want to abort 
>>>>> suspend to RAM due to tsens interrupts since system is already 
>>>>> going into lowest
>>>>> power state. Hence disabling tsens interrupt during suspend to RAM 
>>>>> callback.
>>>>
>>>> Is that a hardware limitation, or a software design choice? I'm not
>>>> sure I want my phone to have thermal notifications disabled when
>>>> it's suspended.
>>>
>>>> Konrad
>>>
>>> As part of suspend to RAM , entire SOC will be off,
>>
>> What do you mean by "entire SOC[sic] will be off"? Surely the memory
>> controller must be on to keep refreshing the memory? Are you thinking
>> of suspend-to-disk (hibernation), by chance?
> 
> Yes, Memory will be in self refreshing  mode(Retained). But SOC will be off
> 
> and will do cold boot to come out of S2R.
> 
>>
>>> this mode (suspend to RAM) is not intended for Mobile product. Tsens 
>>> interrupts are not
>>> disabled as part of suspend to idle(suspend mode for mobile).
>>
>> That's clearly untrue, e.g. the PSCI firmware on SM8550 implements
>> PSCI_SYSTEM_SUSPEND, which does S2R.
> 
> IIUC, PSCI_SYSTEM_SUSPEND will be enabled only for S2R supported 
> products and will be removed it for others.
> 
> Maulik/Raghavendra can comment more
> 
Sorry for delayed response, we have discussed internally on this and
came to the conclusion that disabling tsens interrupt in S2R path is not 
correct approach as S2R is being exercised on mobile kind of products as 
well. I will update the required changes in the next patch.

Priyansh
>>
>> Konrad