On 02/12/2016 02:41 AM, Stephen Boyd wrote:
> On 12/15, Georgi Djakov wrote:
>> diff --git a/drivers/clk/qcom/clk-rpm.c b/drivers/clk/qcom/clk-rpm.c
>> new file mode 100644
>> index 000000000000..7b0e85eefe70
>> --- /dev/null
>> +++ b/drivers/clk/qcom/clk-rpm.c
>> @@ -0,0 +1,290 @@
>> +
>> +static int clk_rpm_prepare(struct clk_hw *hw)
>> +{
>> + struct clk_rpm *r = to_clk_rpm(hw);
>> + unsigned long rate = r->rate;
>> + int ret = 0;
>> +
>> + mutex_lock(&rpm_clk_lock);
>> +
>> + if (!rate)
>> + goto out;
>> +
>> + if (r->branch)
>> + rate = !!rate;
>> +
>> + ret = clk_rpm_set_rate_active(r, rate);
>> +
>> + if (ret)
>> + goto out;
>> +
>> +out:
>> + if (!ret)
>> + r->enabled = true;
>> +
>> + mutex_unlock(&rpm_clk_lock);
>> +
>> + return ret;
>> +}
>
> I don't see any "peer" code in this file. Is there a reason we're
> leaving out the active only vs active + sleep set style clocks?
>
I have left this as a separate follow-up patch, but will squash it
in the next version.
Thanks,
Georgi
