From: Imran Shaik <quic_imrashai@quicinc.com>

Add support to handle the invert logic for branch2 clocks.
Invert branch halt would indicate the clock ON when CLK_OFF
bit is '1' and OFF when CLK_OFF bit is '0'.

Signed-off-by: Imran Shaik <quic_imrashai@quicinc.com>
Signed-off-by: Taniya Das <quic_tdas@quicinc.com>
---
 drivers/clk/qcom/clk-branch.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/clk/qcom/clk-branch.c b/drivers/clk/qcom/clk-branch.c
index f869fc6aaed6..4b24d45be771 100644
--- a/drivers/clk/qcom/clk-branch.c
+++ b/drivers/clk/qcom/clk-branch.c
@@ -48,6 +48,7 @@ static bool clk_branch2_check_halt(const struct clk_branch *br, bool enabling)
 {
 	u32 val;
 	u32 mask;
+	bool invert = (br->halt_check == BRANCH_HALT_ENABLE);
 
 	mask = BRANCH_NOC_FSM_STATUS_MASK << BRANCH_NOC_FSM_STATUS_SHIFT;
 	mask |= BRANCH_CLK_OFF;
@@ -56,9 +57,16 @@ static bool clk_branch2_check_halt(const struct clk_branch *br, bool enabling)
 
 	if (enabling) {
 		val &= mask;
+
+		if (invert)
+			return (val & BRANCH_CLK_OFF) == BRANCH_CLK_OFF;
+
 		return (val & BRANCH_CLK_OFF) == 0 ||
 			val == BRANCH_NOC_FSM_STATUS_ON;
 	} else {
+		if (invert)
+			return (val & BRANCH_CLK_OFF) == 0;
+
 		return val & BRANCH_CLK_OFF;
 	}
 }
-- 
2.17.1

Quoting Taniya Das (2023-04-19 06:30:10)
> From: Imran Shaik <quic_imrashai@quicinc.com>
> 
> Add support to handle the invert logic for branch2 clocks.
> Invert branch halt would indicate the clock ON when CLK_OFF
> bit is '1' and OFF when CLK_OFF bit is '0'.
> 
> Signed-off-by: Imran Shaik <quic_imrashai@quicinc.com>
> Signed-off-by: Taniya Das <quic_tdas@quicinc.com>
> ---
>  drivers/clk/qcom/clk-branch.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/drivers/clk/qcom/clk-branch.c b/drivers/clk/qcom/clk-branch.c
> index f869fc6aaed6..4b24d45be771 100644
> --- a/drivers/clk/qcom/clk-branch.c
> +++ b/drivers/clk/qcom/clk-branch.c
> @@ -48,6 +48,7 @@ static bool clk_branch2_check_halt(const struct clk_branch *br, bool enabling)
>  {
>         u32 val;
>         u32 mask;
> +       bool invert = (br->halt_check == BRANCH_HALT_ENABLE);
>  
>         mask = BRANCH_NOC_FSM_STATUS_MASK << BRANCH_NOC_FSM_STATUS_SHIFT;
>         mask |= BRANCH_CLK_OFF;
> @@ -56,9 +57,16 @@ static bool clk_branch2_check_halt(const struct clk_branch *br, bool enabling)
>  
>         if (enabling) {
>                 val &= mask;
> +
> +               if (invert)
> +                       return (val & BRANCH_CLK_OFF) == BRANCH_CLK_OFF;
> +
>                 return (val & BRANCH_CLK_OFF) == 0 ||
>                         val == BRANCH_NOC_FSM_STATUS_ON;

Do these clks have a NOC_FSM_STATUS bit? I think it would be better to
make a local variable for the val we're looking for, and then test for
that. We may need a mask as well, but the idea is to not duplicate the
test and return from multiple places.

>         } else {
> +               if (invert)
> +                       return (val & BRANCH_CLK_OFF) == 0;
> +
>                 return val & BRANCH_CLK_OFF;
>         }

While at it, I'd get rid of this else and de-indent the code because if
we're 'enabling' we'll return from the function regardless.

Hello Stephen,

Thanks for your review.

On 4/20/2023 3:07 AM, Stephen Boyd wrote:
> Quoting Taniya Das (2023-04-19 06:30:10)
>> From: Imran Shaik <quic_imrashai@quicinc.com>
>>
>> Add support to handle the invert logic for branch2 clocks.
>> Invert branch halt would indicate the clock ON when CLK_OFF
>> bit is '1' and OFF when CLK_OFF bit is '0'.
>>
>> Signed-off-by: Imran Shaik <quic_imrashai@quicinc.com>
>> Signed-off-by: Taniya Das <quic_tdas@quicinc.com>
>> ---
>>   drivers/clk/qcom/clk-branch.c | 8 ++++++++
>>   1 file changed, 8 insertions(+)
>>
>> diff --git a/drivers/clk/qcom/clk-branch.c b/drivers/clk/qcom/clk-branch.c
>> index f869fc6aaed6..4b24d45be771 100644
>> --- a/drivers/clk/qcom/clk-branch.c
>> +++ b/drivers/clk/qcom/clk-branch.c
>> @@ -48,6 +48,7 @@ static bool clk_branch2_check_halt(const struct clk_branch *br, bool enabling)
>>   {
>>          u32 val;
>>          u32 mask;
>> +       bool invert = (br->halt_check == BRANCH_HALT_ENABLE);
>>   
>>          mask = BRANCH_NOC_FSM_STATUS_MASK << BRANCH_NOC_FSM_STATUS_SHIFT;
>>          mask |= BRANCH_CLK_OFF;
>> @@ -56,9 +57,16 @@ static bool clk_branch2_check_halt(const struct clk_branch *br, bool enabling)
>>   
>>          if (enabling) {
>>                  val &= mask;
>> +
>> +               if (invert)
>> +                       return (val & BRANCH_CLK_OFF) == BRANCH_CLK_OFF;
>> +
>>                  return (val & BRANCH_CLK_OFF) == 0 ||
>>                          val == BRANCH_NOC_FSM_STATUS_ON;
> 
> Do these clks have a NOC_FSM_STATUS bit? I think it would be better to
> make a local variable for the val we're looking for, and then test for
> that. We may need a mask as well, but the idea is to not duplicate the
> test and return from multiple places.
> 

Clocks which has invert status doesn't have NOC_FSM_STATUS bit.
Will remove the multiple returns in next patch.

>>          } else {
>> +               if (invert)
>> +                       return (val & BRANCH_CLK_OFF) == 0;
>> +
>>                  return val & BRANCH_CLK_OFF;
>>          }
> 
> While at it, I'd get rid of this else and de-indent the code because if
> we're 'enabling' we'll return from the function regardless.


Yes, Stephen, will take care in the next patch.

-- 
Thanks & Regards,
Taniya Das.