Re: [PATCH 3/3] irqchip/gic-pm: use devm_clk_*() helpers

[Marc Zyngier]

On 06/03/2019 12:00, Jon Hunter wrote:
> 
> On 06/03/2019 11:31, Marc Zyngier wrote:
>> Hi Sameer,
>> 
>> [unrelated to this email: anything that comes from NVIDIA reaches me
>>  encrypted with my public key. Not a big deal, but it'd be good if
>>  someone could fix that.]
>> 
>> On 06/03/2019 10:41, Sameer Pujar wrote:
>>> With pm_clk_*() usage, it is seen that clocks always remain ON. This
>>> happens because clocks are managed by BPMP on Tegra devices and clock
>>> enable/disable happens during prepare/unprepare phase. This patch
>>> avoids use of pm_clk_*() and replaces it with devm_clk_*() helpers.
>>>
>>> Suggested-by: Mohan Kumar D <mkumard@xxxxxxxxxx>
>>> Reviewed-by: Jonathan Hunter <jonathanh@xxxxxxxxxx>
>>> Signed-off-by: Sameer Pujar <spujar@xxxxxxxxxx>
>> 
>> On its own, I'm not opposed to that patch.
>> 
>> But given that there is no in-tree platform using this, despite the code
>> sitting here for more than 2.5 years, this is just updating dead code.
>> 
>> Am I missing anything?
> 
> Nope, but we are working to fix that at long last. I hope in the next
> few months it will not longer be dormant! This driver is still very much
> important to our audio support for newer Tegra devices.

Can we at least make sure it gets compiled when an NVIDIA platform
(ARCH_TEGRA?) is selected? And if this isn't relevant to the current
platforms, then hold off until it actually makes sense.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...