Re: [PATCH v3 1/4] clk: qcom: branch: Extend invert logic for branch2 mem clocks

[Taniya Das]

On 10/24/2025 2:09 PM, Konrad Dybcio wrote:
> On 10/24/25 6:24 AM, Taniya Das wrote:
>> Some clock branches require inverted logic for memory gating, where
>> disabling the memory involves setting a bit and enabling it involves
>> clearing the same bit. This behavior differs from the standard approach
>> memory branch clocks ops where enabling typically sets the bit.
>>
>> Introducing the mem_enable_invert to allow conditional handling of
>> these sequences of the inverted control logic for memory operations
>> required on those memory clock branches.
>>
>> Signed-off-by: Taniya Das <taniya.das@xxxxxxxxxxxxxxxx>
>> ---
>>  drivers/clk/qcom/clk-branch.c | 10 ++++++----
>>  drivers/clk/qcom/clk-branch.h |  4 ++++
>>  2 files changed, 10 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/clk/qcom/clk-branch.c b/drivers/clk/qcom/clk-branch.c
>> index 0f10090d4ae681babbdbbb1b6c68ffe77af7a784..4094ffc53d0c91dfa9e4263134c3a996100ad078 100644
>> --- a/drivers/clk/qcom/clk-branch.c
>> +++ b/drivers/clk/qcom/clk-branch.c
>> @@ -141,9 +141,10 @@ static int clk_branch2_mem_enable(struct clk_hw *hw)
>>  	struct clk_branch branch = mem_br->branch;
>>  	u32 val;
>>  	int ret;
>> +	bool en_val = (mem_br->mem_enable_invert ? false : true);
> 
> " = !mem_br->mem_enable_invert"

will take care in the next patch set.

>>  
>> -	regmap_update_bits(branch.clkr.regmap, mem_br->mem_enable_reg,
>> -			   mem_br->mem_enable_ack_mask, mem_br->mem_enable_ack_mask);
>> +	regmap_assign_bits(branch.clkr.regmap, mem_br->mem_enable_reg,
>> +						mem_br->mem_enable_mask, en_val);
>>  
>>  	ret = regmap_read_poll_timeout(branch.clkr.regmap, mem_br->mem_ack_reg,
>>  				       val, val & mem_br->mem_enable_ack_mask, 0, 200);
>> @@ -158,9 +159,10 @@ static int clk_branch2_mem_enable(struct clk_hw *hw)
>>  static void clk_branch2_mem_disable(struct clk_hw *hw)
>>  {
>>  	struct clk_mem_branch *mem_br = to_clk_mem_branch(hw);
>> +	bool en_val = (mem_br->mem_enable_invert ? true : false);
> 
> This value is a boolean already, you can inline it below
> 

Will fix this in the next patch.

-- 
Thanks,
Taniya Das