Re: [PATCH v6 2/6] drm/etnaviv: add a dedicated function to get various clocks

[Sui Jingfeng]

Hi,

On 2023/6/1 02:07, Lucas Stach wrote:
> > +static int etnaviv_gpu_clk_get(struct etnaviv_gpu *gpu)
> > +{
> > +	struct device *dev = gpu->dev;
> > +
> > +	if (gpu->no_clk)
> > +		return 0;
> > +
> > +	gpu->clk_reg = devm_clk_get_optional(dev, "reg");
> > +	DBG("clk_reg: %p", gpu->clk_reg);
> > +	if (IS_ERR(gpu->clk_reg))
> > +		return PTR_ERR(gpu->clk_reg);
> > +
> > +	gpu->clk_bus = devm_clk_get_optional(dev, "bus");
> > +	DBG("clk_bus: %p", gpu->clk_bus);
> > +	if (IS_ERR(gpu->clk_bus))
> > +		return PTR_ERR(gpu->clk_bus);
> > +
> > +	gpu->clk_core = devm_clk_get(dev, "core");
> > +	DBG("clk_core: %p", gpu->clk_core);
> > +	if (IS_ERR(gpu->clk_core))
> > +		return PTR_ERR(gpu->clk_core);
> > +	gpu->base_rate_core = clk_get_rate(gpu->clk_core);
> > +
> > +	gpu->clk_shader = devm_clk_get_optional(dev, "shader");
> > +	DBG("clk_shader: %p", gpu->clk_shader);
> > +	if (IS_ERR(gpu->clk_shader))
> > +		return PTR_ERR(gpu->clk_shader);
> > +	gpu->base_rate_shader = clk_get_rate(gpu->clk_shader);
> > +
> > +	return 0;
> > +}
> > +
> >   static int etnaviv_gpu_clk_enable(struct etnaviv_gpu *gpu)
> >   {
> >   	int ret;
> >   
> > +	if (gpu->no_clk)
> > +		return 0;
> > +
> I don't see why this would be needed?
I have just tested, this do not needed.
> If your platform doesn't provide
> CONFIG_HAVE_CLK all those functions should be successful no-ops, so
> there is no need to special case this in the driver.

My platform do enable CONFIG_HAVE_CLK,

for ls3a5000 + ls7a1000, my system do not provide device tree support,

that's is to say, there is no DT support.


For ls3a4000 + ls7a1000 platform, the system has DT support, but don't 
has CLK drivers implement toward the clock tree.

typically, our platform's firmware will do such thing(setting a default 
working frequency).


When I first saw etnaviv, I'm also astonishing.

I don't know why there so much clock controllable.

As far as I can understand, my system/hardware have only one clock,

It shall corresponding to the core clk.

> Or does your platform in fact provide a clk subsystem, just the GPU
> clocks are managed by it?
>
> Also all those functions are fine with being called on a NULL clk, so
> shouldn't it be enough to simply avoid calling etnaviv_gpu_clk_get() in
> the PCI device case?
>
> Regards,
> Lucas
--
Jingfeng