Re: [PATCH 2/2] soc: amlogic: ee-pwrc: ensure driver state maches HW state

[Kevin Hilman <khilman@baylibre.com>]

Kevin Hilman <khilman@baylibre.com> writes:

> During init, ensure that the driver on/off state as well as clock
> state matches the hardware state by calling drivers on/off functions
> based on whether the HW state is on/off.
>
> Signed-off-by: Kevin Hilman <khilman@baylibre.com>
> ---
>  drivers/soc/amlogic/meson-ee-pwrc.c | 30 +++++++++--------------------
>  1 file changed, 9 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/soc/amlogic/meson-ee-pwrc.c b/drivers/soc/amlogic/meson-ee-pwrc.c
> index dcce8e694a07..2cb5341aedfa 100644
> --- a/drivers/soc/amlogic/meson-ee-pwrc.c
> +++ b/drivers/soc/amlogic/meson-ee-pwrc.c
> @@ -323,6 +323,8 @@ static int meson_ee_pwrc_init_domain(struct platform_device *pdev,
>  				     struct meson_ee_pwrc *pwrc,
>  				     struct meson_ee_pwrc_domain *dom)
>  {
> +	bool is_off;
> +
>  	dom->pwrc = pwrc;
>  	dom->num_rstc = dom->desc.reset_names_count;
>  	dom->num_clks = dom->desc.clk_names_count;
> @@ -356,27 +358,13 @@ static int meson_ee_pwrc_init_domain(struct platform_device *pdev,
>  	dom->base.power_on = meson_ee_pwrc_on;
>  	dom->base.power_off = meson_ee_pwrc_off;
>  
> -	/*
> -         * TOFIX: This is a special case for the VPU power domain, which can
> -	 * be enabled previously by the bootloader. In this case the VPU
> -         * pipeline may be functional but no driver maybe never attach
> -         * to this power domain, and if the domain is disabled it could
> -         * cause system errors. This is why the pm_domain_always_on_gov
> -         * is used here.
> -         * For the same reason, the clocks should be enabled in case
> -         * we need to power the domain off, otherwise the internal clocks
> -         * prepare/enable counters won't be in sync.
> -         */
> -	if (dom->num_clks && dom->desc.is_off && !dom->desc.is_off(dom)) {
> -		int ret = clk_bulk_prepare_enable(dom->num_clks, dom->clks);
> -		if (ret)
> -			return ret;
> -
> -		pm_genpd_init(&dom->base, &pm_domain_always_on_gov, false);
> -	} else
> -		pm_genpd_init(&dom->base, NULL,
> -			      (dom->desc.is_off ?
> -			       dom->desc.is_off(dom) : true));
> +	/* Ensure that driver state matches HW state */
> +	is_off = dom->desc.is_off ? dom->desc.is_off(dom) : true;
> +	if (is_off)
> +		meson_ee_pwrc_off(&dom->base);

Neil pointed out off-list that this isn't quite right.

This _off() call can potentially try to disable clocks that have never
been enabled (by the clock fwk) resulting in noisy warnings.

I'll send a v2 which always calls _on() and then optionall calls _off().
That will ensure that the drivers notion of the clock state also matches
the HW state.

Kevin