[v1] drm/rockchip: analogix_dp: Apply devm_clk_get_optional() for &rockchip_dp_device.grfclk

[PATCH v1] drm/rockchip: analogix_dp: Apply devm_clk_get_optional() for &rockchip_dp_device.grfclk

Posted by Damon Ding 1 month, 1 week ago

The "grf" clock is optional for Rockchip eDP controller(RK3399 needs
while RK3288 and RK3588 do not).

It can make the code more consice to use devm_clk_get_optional()
instead of devm_clk_get() with extra checks.

Signed-off-by: Damon Ding <damon.ding@rock-chips.com>
---
 drivers/gpu/drm/rockchip/analogix_dp-rockchip.c | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
index d30f0983a53a..d0f78ab9faa6 100644
--- a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
+++ b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
@@ -335,12 +335,8 @@ static int rockchip_dp_of_probe(struct rockchip_dp_device *dp)
 		return PTR_ERR(dp->grf);
 	}
 
-	dp->grfclk = devm_clk_get(dev, "grf");
-	if (PTR_ERR(dp->grfclk) == -ENOENT) {
-		dp->grfclk = NULL;
-	} else if (PTR_ERR(dp->grfclk) == -EPROBE_DEFER) {
-		return -EPROBE_DEFER;
-	} else if (IS_ERR(dp->grfclk)) {
+	dp->grfclk = devm_clk_get_optional(dev, "grf");
+	if (IS_ERR(dp->grfclk)) {
 		DRM_DEV_ERROR(dev, "failed to get grf clock\n");
 		return PTR_ERR(dp->grfclk);
 	}
-- 
2.34.1

Re: [PATCH v1] drm/rockchip: analogix_dp: Apply devm_clk_get_optional() for &rockchip_dp_device.grfclk

Posted by Heiko Stuebner 1 month, 1 week ago

Hi Damon,

Am Sonntag, 24. August 2025, 05:19:32 Mitteleuropäische Sommerzeit schrieb Damon Ding:
> The "grf" clock is optional for Rockchip eDP controller(RK3399 needs
> while RK3288 and RK3588 do not).
> 
> It can make the code more consice to use devm_clk_get_optional()
> instead of devm_clk_get() with extra checks.
> 
> Signed-off-by: Damon Ding <damon.ding@rock-chips.com>
> ---
>  drivers/gpu/drm/rockchip/analogix_dp-rockchip.c | 8 ++------
>  1 file changed, 2 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
> index d30f0983a53a..d0f78ab9faa6 100644
> --- a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
> +++ b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
> @@ -335,12 +335,8 @@ static int rockchip_dp_of_probe(struct rockchip_dp_device *dp)
>  		return PTR_ERR(dp->grf);
>  	}
>  
> -	dp->grfclk = devm_clk_get(dev, "grf");
> -	if (PTR_ERR(dp->grfclk) == -ENOENT) {
> -		dp->grfclk = NULL;
> -	} else if (PTR_ERR(dp->grfclk) == -EPROBE_DEFER) {
> -		return -EPROBE_DEFER;
> -	} else if (IS_ERR(dp->grfclk)) {
> +	dp->grfclk = devm_clk_get_optional(dev, "grf");
> +	if (IS_ERR(dp->grfclk)) {
>  		DRM_DEV_ERROR(dev, "failed to get grf clock\n");
>  		return PTR_ERR(dp->grfclk);

generall I definitly I agree, but you're loosing the
"don't print anything for EPROBE_DEFER" handling.

Additionally DRM_DEV_ERROR says in the code
"this is deprecated in favor of drm_err() or dev_err()." .

Also this in the probe-path, so to catch both issues, I'd sugest
using dev_err_probe() for the message, which handles EPROBE_DEFER
automatically.

Heiko

Re: [PATCH v1] drm/rockchip: analogix_dp: Apply devm_clk_get_optional() for &rockchip_dp_device.grfclk

Posted by Damon Ding 1 month, 1 week ago

Hi Heiko,

On 8/24/2025 2:08 PM, Heiko Stuebner wrote:
> Hi Damon,
> 
> Am Sonntag, 24. August 2025, 05:19:32 Mitteleuropäische Sommerzeit schrieb Damon Ding:
>> The "grf" clock is optional for Rockchip eDP controller(RK3399 needs
>> while RK3288 and RK3588 do not).
>>
>> It can make the code more consice to use devm_clk_get_optional()
>> instead of devm_clk_get() with extra checks.
>>
>> Signed-off-by: Damon Ding <damon.ding@rock-chips.com>
>> ---
>>   drivers/gpu/drm/rockchip/analogix_dp-rockchip.c | 8 ++------
>>   1 file changed, 2 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
>> index d30f0983a53a..d0f78ab9faa6 100644
>> --- a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
>> +++ b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c
>> @@ -335,12 +335,8 @@ static int rockchip_dp_of_probe(struct rockchip_dp_device *dp)
>>   		return PTR_ERR(dp->grf);
>>   	}
>>   
>> -	dp->grfclk = devm_clk_get(dev, "grf");
>> -	if (PTR_ERR(dp->grfclk) == -ENOENT) {
>> -		dp->grfclk = NULL;
>> -	} else if (PTR_ERR(dp->grfclk) == -EPROBE_DEFER) {
>> -		return -EPROBE_DEFER;
>> -	} else if (IS_ERR(dp->grfclk)) {
>> +	dp->grfclk = devm_clk_get_optional(dev, "grf");
>> +	if (IS_ERR(dp->grfclk)) {
>>   		DRM_DEV_ERROR(dev, "failed to get grf clock\n");
>>   		return PTR_ERR(dp->grfclk);
> 
> generall I definitly I agree, but you're loosing the
> "don't print anything for EPROBE_DEFER" handling.
> 
> Additionally DRM_DEV_ERROR says in the code
> "this is deprecated in favor of drm_err() or dev_err()." .
> 
> Also this in the probe-path, so to catch both issues, I'd sugest
> using dev_err_probe() for the message, which handles EPROBE_DEFER
> automatically.
> 

Yeah, I will use dev_err_probe() instead in the next version.

In addition, it should be nice to modify the other similar codes in 
rockchip_dp_of_probe(). I will add another commit to do it.

Best regards,
Damon