lkml.org 
[lkml]   [2023]   [Jan]   [24]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH 2/7] clk: qcom: Add Global Clock Controller driver for IPQ9574
From


On 1/24/2023 3:23 PM, Konrad Dybcio wrote:
>
>
> On 24.01.2023 08:27, Devi Priya wrote:
>>
>>
>> On 1/13/2023 7:39 PM, Konrad Dybcio wrote:
>>>
>>>
>>> On 13.01.2023 14:21, Devi Priya wrote:
>>>>
>>>>
>>>> On 1/10/2023 6:07 PM, Konrad Dybcio wrote:
>>>>>
>>>>>
>>>>> On 10.01.2023 13:13, devi priya wrote:
>>>>>> Add Global Clock Controller (GCC) driver for ipq9574 based devices
>>>>>>
>>>>>> Co-developed-by: Anusha Rao <quic_anusha@quicinc.com>
>>>>>> Signed-off-by: Anusha Rao <quic_anusha@quicinc.com>
>>>>>> Signed-off-by: devi priya <quic_devipriy@quicinc.com>
>>>>>> ---
>>> [...]
>>>
>>>>>> +static struct clk_branch gcc_blsp1_qup6_i2c_apps_clk = {
>>>>>> +    .halt_reg = 0x07024,
>>>>>> +    .clkr = {
>>>>>> +        .enable_reg = 0x07024,
>>>>>> +        .enable_mask = BIT(0),
>>>>>> +        .hw.init = &(struct clk_init_data) {
>>>>>> +            .name = "gcc_blsp1_qup6_i2c_apps_clk",
>>>>>> +            .parent_hws = (const struct clk_hw *[]) {
>>>>>> +                    &blsp1_qup6_i2c_apps_clk_src.clkr.hw },
>>>>>> +            .num_parents = 1,
>>>>>> +            .flags = CLK_SET_RATE_PARENT | CLK_IGNORE_UNUSED,
>>>>> Sounds very much like a hack..
>>>> Got it, will remove the clock entry as it is not being used in linux
>>> I'm not sure removing it is the best option, somebody might have a
>>> funky board where they use this particular QUP for I2C for whatever
>>> reason and then the clock would have to be re-added..
>> Sure, Understood
>> This clock is used by the RPM component to communicate with PMIC and we
>> would add the critical flag here
> Okay, so this SoC is intended to ship with some RPM PMICs and
> *always* with an I2C companion that's required for some basic
> functionality, correct?
>
> Otherwise, if it's just for wifi/multimedia/etc (like PM8008ij
> on some newer devices), you should not make it critical and
> simply rely on Linux keeping it alive like so:
>
> consumer takes a regulator
> the regulator does not go to sleep because it's consumed
> the PMIC is active because a regulator on it is being used
> the I2C bus is active because its child PMIC is used
> the I2C clocks are alive because there's an active user
>
Yes correct, the SoC would always have an I2C companion
> Konrad
>>
>>> Thanks for addressing all of the review comments so thoroughly!
>>>
>>> Konrad
>>
>> Best Regards,
>> Devi Priya

\
 
 \ /
  Last update: 2023-03-26 23:54    [W:0.130 / U:0.300 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site