On 06/05, Sudeep Holla wrote:

On 05/06/15 10:36, Sudeep Holla wrote:

quoted

On 04/06/15 21:20, Stephen Boyd wrote:

quoted

On 05/27, Sudeep Holla wrote:

[...]

quoted

quoted

quoted

+
+static unsigned long scpi_clk_recalc_rate(struct clk_hw *hw,
+					  unsigned long parent_rate)
+{
+	struct scpi_clk *clk = to_scpi_clk(hw);
+
+	return scpi_ops->clk_get_val(clk->id);
+}
+
+static long scpi_clk_round_rate(struct clk_hw *hw, unsigned long rate,
+				unsigned long *parent_rate)
+{
+	struct scpi_clk *clk = to_scpi_clk(hw);
+
+	if (WARN_ON(clk->rate_min && rate < clk->rate_min))
+		rate = clk->rate_min;
+	if (WARN_ON(clk->rate_max && rate > clk->rate_max))
+		rate = clk->rate_max;
+
+	return rate;
+}

Hm.. this seems really generic. It might be better to support a
way to tell the framework to limit the min/max rate that's
accepted for a clk. That could be done later though.

True, framework have some boundary checks in place. I will check if
I can use it with minimum changes to the core. If not, we can take this
up later as you suggested.

I found that the framework already provides clk_set_rate_range for this
purpose. Sorry for missing this earlier(seems like that's added quite
recently in v4.0). I think I still need to retain round_rate as the core
framework insists.

Sure, or use determine_rate if you want to limit the min/max from
the clk provider itself.

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project