On 9/1/21 5:37 AM, Avri Altman wrote:
> It is essentially up to the platform to decide what further actions need
> to be taken. So add a designated vop for that. Each chipset vendor can
> decide if it wants to use the thermal subsystem, hw monitor, or some
> Privet implementation.
Why to make chipset vendors define what to do in case of extreme
temperatures? I'd prefer a single implementation in ufshcd.c instead of
making each vendor come up with a different implementation.
> + void (*temp_notify)(struct ufs_hba *hba, u16 status);
Please do not add new vops without adding at least one implementation of
that vop.
Thanks,
Bart.
On 9/1/2021 9:39 AM, Bart Van Assche wrote:
> On 9/1/21 5:37 AM, Avri Altman wrote:
>> It is essentially up to the platform to decide what further actions need
>> to be taken. So add a designated vop for that. Each chipset vendor can
>> decide if it wants to use the thermal subsystem, hw monitor, or some
>> Privet implementation.
>
> Why to make chipset vendors define what to do in case of extreme
> temperatures? I'd prefer a single implementation in ufshcd.c instead of
> making each vendor come up with a different implementation.
>
I think it should be either i.e. if a vendor specific implementation is
defined use that else use the generic implementation in ufshcd.
There may be a bunch of things that each vendor may need/want do
depending upon use-case, I imagine.
>> + void (*temp_notify)(struct ufs_hba *hba, u16 status);
>
> Please do not add new vops without adding at least one implementation of
> that vop.
>
> Thanks,
>
> Bart.
--
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
Linux Foundation Collaborative Project
> On 9/1/2021 9:39 AM, Bart Van Assche wrote:
> > On 9/1/21 5:37 AM, Avri Altman wrote:
> >> It is essentially up to the platform to decide what further actions need
> >> to be taken. So add a designated vop for that. Each chipset vendor can
> >> decide if it wants to use the thermal subsystem, hw monitor, or some
> >> Privet implementation.
> >
> > Why to make chipset vendors define what to do in case of extreme
> > temperatures? I'd prefer a single implementation in ufshcd.c instead of
> > making each vendor come up with a different implementation.
The storage device is merely acting as a temperature sensor.
This info, jointly with other temperature sensors of the system,
Should be used elsewhere in a much broader scope - probably by Android.
Either way, ufshcd is hardly the place for those decisions.
> >
> I think it should be either i.e. if a vendor specific implementation is
> defined use that else use the generic implementation in ufshcd.
> There may be a bunch of things that each vendor may need/want do
> depending upon use-case, I imagine.
I agree, and this is why I wanted to allow that that flexibility.
But I get Bart's point. I will register the sensor in some subsystem.
It should allow the required degrees of freedom.
Thanks,
Avri
>
> >> + void (*temp_notify)(struct ufs_hba *hba, u16 status);
> >
> > Please do not add new vops without adding at least one implementation of
> > that vop.
> >
> > Thanks,
> >
> > Bart.
>
>
> --
> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
> Linux Foundation Collaborative Project