Re: [PATCH/RFT] soc: amlogic: meson-gx-pwrc-vpu: switch to clk_bulk

[Kevin Hilman <khilman@baylibre.com>]

Neil Armstrong <narmstrong@baylibre.com> writes:

> Hi,
>
> On 14/08/2019 00:47, Kevin Hilman wrote:
>> Hi Neil,
>> 
>> Kevin Hilman <khilman@baylibre.com> writes:
>> 
>>> Instead of expecting a specific number of clocks with specific clock
>>> names, switch to using the bulk clock API.
>>>
>>> This is a first step towards generalizing this driver to work with
>>> other domains.
>>>
>>> Cc: Neil Armstrong <narmstrong@baylibre.com>
>>> Signed-off-by: Kevin Hilman <khilman@baylibre.com>
>>> ---
>>> Boot tested on meson-g12a-sei510 and verified that framebuffer console
>>> comes up and still works.
>>>
>>>  drivers/soc/amlogic/meson-gx-pwrc-vpu.c | 41 ++++++-------------------
>>>  1 file changed, 10 insertions(+), 31 deletions(-)
>>>
>>> diff --git a/drivers/soc/amlogic/meson-gx-pwrc-vpu.c b/drivers/soc/amlogic/meson-gx-pwrc-vpu.c
>>> index 511b6856225d..5f6519f43a31 100644
>>> --- a/drivers/soc/amlogic/meson-gx-pwrc-vpu.c
>>> +++ b/drivers/soc/amlogic/meson-gx-pwrc-vpu.c
>>> @@ -34,8 +34,8 @@ struct meson_gx_pwrc_vpu {
>>>  	struct regmap *regmap_ao;
>>>  	struct regmap *regmap_hhi;
>>>  	struct reset_control *rstc;
>>> -	struct clk *vpu_clk;
>>> -	struct clk *vapb_clk;
>>> +	struct clk_bulk_data *clks;
>>> +	int num_clks;
>>>  };
>>>  
>>>  static inline
>>> @@ -76,8 +76,7 @@ static int meson_gx_pwrc_vpu_power_off(struct generic_pm_domain *genpd)
>>>  
>>>  	msleep(20);
>>>  
>>> -	clk_disable_unprepare(pd->vpu_clk);
>>> -	clk_disable_unprepare(pd->vapb_clk);
>>> +	clk_bulk_disable_unprepare(pd->num_clks, pd->clks);
>> 
>> Note the original turn-off order here is VPU then VAPB...
>> 
>>>  	return 0;
>>>  }
>>> @@ -119,25 +118,14 @@ static int meson_g12a_pwrc_vpu_power_off(struct generic_pm_domain *genpd)
>>>  
>>>  	msleep(20);
>>>  
>>> -	clk_disable_unprepare(pd->vpu_clk);
>>> -	clk_disable_unprepare(pd->vapb_clk);
>>> +	clk_bulk_disable_unprepare(pd->num_clks, pd->clks);
>> 
>> ... and the origianl turn-on ordr is also VPU then VAPB.
>> 
>> Using the clock bulk API, the new turn-on order will be the order they
>> clocks appear in DT.  The turn-off order will be the reverse of that.
>> 
>> That seems right to me, but it is a change in behavior from the current
>> code.
>> 
>> Did you set the enable and disable ordering the same for any specific
>> reason?  Any reason to thing reversing the disable order is going to
>> cause any issues?
>
> No the order is not an issue here, the 2 clocks feeds 2 different parts of the VPU,
> one is the APB register bridge (vapb) and the other feeds the vpu video pipeline,
> so the order is not an issue.

OK, thanks for confirming.

Kevin

_______________________________________________
linux-amlogic mailing list
linux-amlogic@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-amlogic