Re: [RESEND PATCH 4/8] clk: core: support clocks which requires parents on (part 2)

From: Michael Turquette
Date: Fri Jun 24 2016 - 19:53:08 EST


Hi Dong,

Quoting Dong Aisheng (2016-04-20 02:34:36)
> On Freescale i.MX7D platform, all clocks operations, including
> enable/disable, rate change and re-parent, requires its parent clock on.
> Current clock core can not support it well.
> This patch adding flag CLK_OPS_PARENT_ON to handle this special case in
> clock core that enable its parent clock firstly for each operation and
> disable it later after operation complete.
>
> The patch part 2 fixes set clock rate and set parent while its parent
> is off. The most special case is for set_parent() operation which requires
> all parents including both old and new one to be enabled at the same time
> during the operation.
>
> Cc: Michael Turquette <mturquette@xxxxxxxxxxxx>
> Cc: Stephen Boyd <sboyd@xxxxxxxxxxxxxx>
> Cc: Shawn Guo <shawnguo@xxxxxxxxxx>
> Signed-off-by: Dong Aisheng <aisheng.dong@xxxxxxx>
> ---
> drivers/clk/clk.c | 45 ++++++++++++++++++++++++++++-----------------
> 1 file changed, 28 insertions(+), 17 deletions(-)
>
> diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
> index 1f054d1a..d6cef61 100644
> --- a/drivers/clk/clk.c
> +++ b/drivers/clk/clk.c
> @@ -1174,7 +1174,7 @@ static struct clk_core *__clk_set_parent_before(struct clk_core *core,
> struct clk_core *old_parent = core->parent;
>
> /*
> - * Migrate prepare state between parents and prevent race with
> + * 1. Migrate prepare state between parents and prevent race with
> * clk_enable().
> *
> * If the clock is not prepared, then a race with
> @@ -1189,13 +1189,16 @@ static struct clk_core *__clk_set_parent_before(struct clk_core *core,
> * hardware and software states.
> *
> * See also: Comment for clk_set_parent() below.
> + *
> + * 2. enable parent clocks for CLK_OPS_PARENT_ON clock
> */
> - if (core->prepare_count) {
> - clk_core_prepare(parent);
> - flags = clk_enable_lock();
> - clk_core_enable(parent);
> - clk_core_enable(core);
> - clk_enable_unlock(flags);
> + if (core->prepare_count || core->flags & CLK_OPS_PARENT_ON) {
> + clk_core_prepare_enable(parent);
> + if (core->prepare_count)
> + clk_core_enable_lock(core);
> + else
> + clk_core_prepare_enable(old_parent);

It took me a few seconds to figure out what I was looking at. The
set_parent stuff is tricky enough already ... I think we should keep the
state machine logic very simple. How about:

/* enable old_parent & parent if CLK_OPS_PARENT_ENABLE is set */
if (core->flags & CLK_OPS_PARENT_ENABLE) {
clk_core_prepare_enable(old_parent);
clk_core_prepare_enable(parent);
}

/* migrate prepare count if > 0 */
if (core->prepare_count) {
clk_core_prepare_enable(parent);
clk_core_enable_lock(core);
}

> +
> }
>
> /* update the clk tree topology */
> @@ -1210,18 +1213,16 @@ static void __clk_set_parent_after(struct clk_core *core,
> struct clk_core *parent,
> struct clk_core *old_parent)
> {
> - unsigned long flags;
> -
> /*
> * Finish the migration of prepare state and undo the changes done
> * for preventing a race with clk_enable().
> */
> - if (core->prepare_count) {
> - flags = clk_enable_lock();
> - clk_core_disable(core);
> - clk_core_disable(old_parent);
> - clk_enable_unlock(flags);
> - clk_core_unprepare(old_parent);
> + if (core->prepare_count || core->flags & CLK_OPS_PARENT_ON) {
> + clk_core_disable_unprepare(old_parent);
> + if (core->prepare_count)
> + clk_core_disable_lock(core);
> + else
> + clk_core_disable_unprepare(parent);

Similarly, a more readable solution is:

/*
* Finish the migration of prepare state and undo the changes done
* for preventing a race with clk_enable().
*/
if (core->prepare_count) {
clk_core_disable_lock(core);
clk_core_disable_unprepare(old_parent);

/* re-balance ref counting if CLK_OPS_PARENT_ENABLE is set */
if (core->flags & CLK_OPS_PARENT_ENABLE) {
clk_core_disable_unprepare(parent);
clk_core_disable_unprepare(old_parent);
}

> }
> }
>
> @@ -1468,13 +1469,17 @@ static void clk_change_rate(struct clk_core *core)
> unsigned long best_parent_rate = 0;
> bool skip_set_rate = false;
> struct clk_core *old_parent;
> + struct clk_core *parent = NULL;
>
> old_rate = core->rate;
>
> - if (core->new_parent)
> + if (core->new_parent) {
> + parent = core->new_parent;
> best_parent_rate = core->new_parent->rate;
> - else if (core->parent)
> + } else if (core->parent) {
> + parent = core->parent;
> best_parent_rate = core->parent->rate;
> + }
>
> if (core->flags & CLK_SET_RATE_UNGATE) {
> unsigned long flags;
> @@ -1504,6 +1509,9 @@ static void clk_change_rate(struct clk_core *core)
>
> trace_clk_set_rate(core, core->new_rate);
>
> + if (core->flags & CLK_OPS_PARENT_ON)
> + clk_core_prepare_enable(parent);
> +
> if (!skip_set_rate && core->ops->set_rate)
> core->ops->set_rate(core->hw, core->new_rate, best_parent_rate);
>
> @@ -1520,6 +1528,9 @@ static void clk_change_rate(struct clk_core *core)
> clk_core_unprepare(core);
> }
>
> + if (core->flags & CLK_OPS_PARENT_ON)
> + clk_core_disable_unprepare(parent);
> +

This part looks good to me.

Regards,
Mike

> if (core->notifier_count && old_rate != core->rate)
> __clk_notify(core, POST_RATE_CHANGE, old_rate, core->rate);
>
> --
> 1.9.1
>