From: Alice Guo <alice.guo@nxp.com>

i.MX93 watchdog needn't wait 2.5 clocks after RCS is done. So set
post_rcs_wait to false for "fsl,imx93-wdt".

Signed-off-by: Alice Guo <alice.guo@nxp.com>
Reviewed-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Frank Li <Frank.Li@nxp.com>
---
 drivers/watchdog/imx7ulp_wdt.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/watchdog/imx7ulp_wdt.c b/drivers/watchdog/imx7ulp_wdt.c
index 904b9f1873856..3a75a6f98f8f0 100644
--- a/drivers/watchdog/imx7ulp_wdt.c
+++ b/drivers/watchdog/imx7ulp_wdt.c
@@ -405,7 +405,6 @@ static const struct imx_wdt_hw_feature imx8ulp_wdt_hw = {
 static const struct imx_wdt_hw_feature imx93_wdt_hw = {
 	.prescaler_enable = true,
 	.wdog_clock_rate = 125,
-	.post_rcs_wait = true,
 };
 
 static const struct of_device_id imx7ulp_wdt_dt_ids[] = {

-- 
2.34.1

On 7/11/24 15:41, Frank Li wrote:
> From: Alice Guo <alice.guo@nxp.com>
> 
> i.MX93 watchdog needn't wait 2.5 clocks after RCS is done. So set
> post_rcs_wait to false for "fsl,imx93-wdt".
> 
> Signed-off-by: Alice Guo <alice.guo@nxp.com>
> Reviewed-by: Ye Li <ye.li@nxp.com>
> Signed-off-by: Frank Li <Frank.Li@nxp.com>
> ---
>   drivers/watchdog/imx7ulp_wdt.c | 1 -
>   1 file changed, 1 deletion(-)
> 
> diff --git a/drivers/watchdog/imx7ulp_wdt.c b/drivers/watchdog/imx7ulp_wdt.c
> index 904b9f1873856..3a75a6f98f8f0 100644
> --- a/drivers/watchdog/imx7ulp_wdt.c
> +++ b/drivers/watchdog/imx7ulp_wdt.c
> @@ -405,7 +405,6 @@ static const struct imx_wdt_hw_feature imx8ulp_wdt_hw = {
>   static const struct imx_wdt_hw_feature imx93_wdt_hw = {
>   	.prescaler_enable = true,
>   	.wdog_clock_rate = 125,
> -	.post_rcs_wait = true,
>   };
>   
>   static const struct of_device_id imx7ulp_wdt_dt_ids[] = {
> 
Introducing that flag in the previous patch just to remove it here doesn't
make sense to me, sorry.

What the two changes do together is to disable post_rcs_wait for iMX93.
That is a single logical change, and it can and should be done in a
single patch. If you do that by moving the flag into imx_wdt_hw_feature
or by adding another of_device_is_compatible() is your call.

Guenter

On Thu, Jul 11, 2024 at 03:55:52PM -0700, Guenter Roeck wrote:
> On 7/11/24 15:41, Frank Li wrote:
> > From: Alice Guo <alice.guo@nxp.com>
> > 
> > i.MX93 watchdog needn't wait 2.5 clocks after RCS is done. So set
> > post_rcs_wait to false for "fsl,imx93-wdt".
> > 
> > Signed-off-by: Alice Guo <alice.guo@nxp.com>
> > Reviewed-by: Ye Li <ye.li@nxp.com>
> > Signed-off-by: Frank Li <Frank.Li@nxp.com>
> > ---
> >   drivers/watchdog/imx7ulp_wdt.c | 1 -
> >   1 file changed, 1 deletion(-)
> > 
> > diff --git a/drivers/watchdog/imx7ulp_wdt.c b/drivers/watchdog/imx7ulp_wdt.c
> > index 904b9f1873856..3a75a6f98f8f0 100644
> > --- a/drivers/watchdog/imx7ulp_wdt.c
> > +++ b/drivers/watchdog/imx7ulp_wdt.c
> > @@ -405,7 +405,6 @@ static const struct imx_wdt_hw_feature imx8ulp_wdt_hw = {
> >   static const struct imx_wdt_hw_feature imx93_wdt_hw = {
> >   	.prescaler_enable = true,
> >   	.wdog_clock_rate = 125,
> > -	.post_rcs_wait = true,
> >   };
> >   static const struct of_device_id imx7ulp_wdt_dt_ids[] = {
> > 
> Introducing that flag in the previous patch just to remove it here doesn't
> make sense to me, sorry.

Some maintainer want create function equal patch first if just code
restructure/re-originzed. Then add additional change base on it.

Of course, I can squash to one if you like.

Frank
 
> 
> What the two changes do together is to disable post_rcs_wait for iMX93.
> That is a single logical change, and it can and should be done in a
> single patch. If you do that by moving the flag into imx_wdt_hw_feature
> or by adding another of_device_is_compatible() is your call.
> 
> Guenter
>

On 7/11/24 18:39, Frank Li wrote:
> On Thu, Jul 11, 2024 at 03:55:52PM -0700, Guenter Roeck wrote:
>> On 7/11/24 15:41, Frank Li wrote:
>>> From: Alice Guo <alice.guo@nxp.com>
>>>
>>> i.MX93 watchdog needn't wait 2.5 clocks after RCS is done. So set
>>> post_rcs_wait to false for "fsl,imx93-wdt".
>>>
>>> Signed-off-by: Alice Guo <alice.guo@nxp.com>
>>> Reviewed-by: Ye Li <ye.li@nxp.com>
>>> Signed-off-by: Frank Li <Frank.Li@nxp.com>
>>> ---
>>>    drivers/watchdog/imx7ulp_wdt.c | 1 -
>>>    1 file changed, 1 deletion(-)
>>>
>>> diff --git a/drivers/watchdog/imx7ulp_wdt.c b/drivers/watchdog/imx7ulp_wdt.c
>>> index 904b9f1873856..3a75a6f98f8f0 100644
>>> --- a/drivers/watchdog/imx7ulp_wdt.c
>>> +++ b/drivers/watchdog/imx7ulp_wdt.c
>>> @@ -405,7 +405,6 @@ static const struct imx_wdt_hw_feature imx8ulp_wdt_hw = {
>>>    static const struct imx_wdt_hw_feature imx93_wdt_hw = {
>>>    	.prescaler_enable = true,
>>>    	.wdog_clock_rate = 125,
>>> -	.post_rcs_wait = true,
>>>    };
>>>    static const struct of_device_id imx7ulp_wdt_dt_ids[] = {
>>>
>> Introducing that flag in the previous patch just to remove it here doesn't
>> make sense to me, sorry.
> 
> Some maintainer want create function equal patch first if just code
> restructure/re-originzed. Then add additional change base on it.
> 

In general I would ask you to do that as well, but not if patch 1/2 introduces
a change and patch 2/2 does nothing but to remove part of the change introduced
in patch 1/2.

Guenter