Re: [PATCH v3 06/15] clk: tegra: Remove tegra_pmc_clk_init along with clk ids

From: Dmitry Osipenko
Date: Sat Dec 07 2019 - 10:04:23 EST


07.12.2019 17:43, Dmitry Osipenko ÐÐÑÐÑ:
> 07.12.2019 17:33, Dmitry Osipenko ÐÐÑÐÑ:
>> 06.12.2019 05:48, Sowjanya Komatineni ÐÐÑÐÑ:
>>> Current Tegra clock driver registers PMC clocks clk_out_1, clk_out_2,
>>> clk_out_3 and blink output in tegra_pmc_init() which does direct Tegra
>>> PMC access during clk_ops and these PMC register read and write access
>>> will not happen when PMC is in secure mode.
>>>
>>> Any direct PMC register access from non-secure world will not go
>>> through and all the PMC clocks and blink control are done in Tegra PMC
>>> driver with PMC as clock provider.
>>>
>>> This patch removes tegra_pmc_clk_init along with corresponding clk ids
>>> from Tegra clock driver.
>>>
>>> Signed-off-by: Sowjanya Komatineni <skomatineni@xxxxxxxxxx>
>>> ---
>>
>> [snip]
>>
>>> @@ -1230,9 +1222,6 @@ static struct tegra_clk_init_table init_table[] __initdata = {
>>
>>> { TEGRA30_CLK_PLL_A, TEGRA30_CLK_CLK_MAX, 564480000, 1 },
>>> { TEGRA30_CLK_PLL_A_OUT0, TEGRA30_CLK_CLK_MAX, 11289600, 1 },
>>> { TEGRA30_CLK_EXTERN1, TEGRA30_CLK_PLL_A_OUT0, 0, 1 },
>>
>> Perhaps these clocks do not need to be always-enabled?
>>
>> [snip]
>>
>
> Also, EXTERN1 parent configuration should be moved to the audio
> driver/device-tree as well.

Ah, I missed that it's done in the patch #10.

> Maybe it even makes sense to move the whole configuration, including
> PLLA. I don't see why clk driver need to do something for the audio driver.