i2c: qcom-geni: Simplify PM resume error handling

[PATCH] i2c: qcom-geni: Simplify PM resume error handling

Posted by Mukesh Kumar Savaliya 1 month, 3 weeks ago

The driver currently logs pm_runtime_resume_and_get() failures with
dev_err() and returns the error code separately.

Replace this with dev_err_probe(), which combines the dev_err() log
message and the return value propagation into a single call, consistent
with how all other error paths in this driver already report failures.

No functional change intended.

Signed-off-by: Mukesh Kumar Savaliya <mukesh.savaliya@oss.qualcomm.com>
---
 drivers/i2c/busses/i2c-qcom-geni.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/i2c-qcom-geni.c
index d5c977e34994..96015e077b69 100644
--- a/drivers/i2c/busses/i2c-qcom-geni.c
+++ b/drivers/i2c/busses/i2c-qcom-geni.c
@@ -994,10 +994,8 @@ static int geni_i2c_init(struct geni_i2c_dev *gi2c)
 	int ret;
 
 	ret = pm_runtime_resume_and_get(gi2c->se.dev);
-	if (ret < 0) {
-		dev_err(gi2c->se.dev, "error turning on device :%d\n", ret);
-		return ret;
-	}
+	if (ret < 0)
+		return dev_err_probe(gi2c->se.dev, ret, "error turning on device\n");
 
 	proto = geni_se_read_proto(&gi2c->se);
 	if (proto == GENI_SE_INVALID_PROTO) {
-- 
2.43.0

Re: [PATCH] i2c: qcom-geni: Simplify PM resume error handling

Posted by Andi Shyti 3 weeks ago

Hi Mukesh,

On Wed, Aug 05, 2026 at 12:30:13PM +0530, Mukesh Kumar Savaliya wrote:
> The driver currently logs pm_runtime_resume_and_get() failures with
> dev_err() and returns the error code separately.
> 
> Replace this with dev_err_probe(), which combines the dev_err() log
> message and the return value propagation into a single call, consistent
> with how all other error paths in this driver already report failures.
> 
> No functional change intended.
> 
> Signed-off-by: Mukesh Kumar Savaliya <mukesh.savaliya@oss.qualcomm.com>

merged to i2c/i2c.

Thanks,
Andi

Re: [PATCH] i2c: qcom-geni: Simplify PM resume error handling

Posted by Viken Dadhaniya 1 month, 3 weeks ago


On 8/5/2026 12:30 PM, Mukesh Kumar Savaliya wrote:
> The driver currently logs pm_runtime_resume_and_get() failures with
> dev_err() and returns the error code separately.
> 
> Replace this with dev_err_probe(), which combines the dev_err() log
> message and the return value propagation into a single call, consistent
> with how all other error paths in this driver already report failures.
> 
> No functional change intended.
> 
> Signed-off-by: Mukesh Kumar Savaliya <mukesh.savaliya@oss.qualcomm.com>
> ---
>  drivers/i2c/busses/i2c-qcom-geni.c | 6 ++----
>  1 file changed, 2 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/i2c-qcom-geni.c
> index d5c977e34994..96015e077b69 100644
> --- a/drivers/i2c/busses/i2c-qcom-geni.c
> +++ b/drivers/i2c/busses/i2c-qcom-geni.c
> @@ -994,10 +994,8 @@ static int geni_i2c_init(struct geni_i2c_dev *gi2c)
>  	int ret;
>  
>  	ret = pm_runtime_resume_and_get(gi2c->se.dev);
> -	if (ret < 0) {
> -		dev_err(gi2c->se.dev, "error turning on device :%d\n", ret);
> -		return ret;
> -	}
> +	if (ret < 0)
> +		return dev_err_probe(gi2c->se.dev, ret, "error turning on device\n");
>  
>  	proto = geni_se_read_proto(&gi2c->se);
>  	if (proto == GENI_SE_INVALID_PROTO) {

Reviewed-by: Viken Dadhaniya <viken.dadhaniya@oss.qualcomm.com>

Re: [PATCH] i2c: qcom-geni: Simplify PM resume error handling

Posted by Praveen Talari 1 month, 3 weeks ago

On 05-08-2026 12:30, Mukesh Kumar Savaliya wrote:
> The driver currently logs pm_runtime_resume_and_get() failures with
> dev_err() and returns the error code separately.
>
> Replace this with dev_err_probe(), which combines the dev_err() log
> message and the return value propagation into a single call, consistent
> with how all other error paths in this driver already report failures.
>
> No functional change intended.
>
> Signed-off-by: Mukesh Kumar Savaliya <mukesh.savaliya@oss.qualcomm.com>
> ---
>   drivers/i2c/busses/i2c-qcom-geni.c | 6 ++----
>   1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/i2c-qcom-geni.c
> index d5c977e34994..96015e077b69 100644
> --- a/drivers/i2c/busses/i2c-qcom-geni.c
> +++ b/drivers/i2c/busses/i2c-qcom-geni.c
> @@ -994,10 +994,8 @@ static int geni_i2c_init(struct geni_i2c_dev *gi2c)
>   	int ret;
>   
>   	ret = pm_runtime_resume_and_get(gi2c->se.dev);
> -	if (ret < 0) {
> -		dev_err(gi2c->se.dev, "error turning on device :%d\n", ret);
> -		return ret;
> -	}
> +	if (ret < 0)
> +		return dev_err_probe(gi2c->se.dev, ret, "error turning on device\n");
>   
>   	proto = geni_se_read_proto(&gi2c->se);
>   	if (proto == GENI_SE_INVALID_PROTO) {
Acked-by: Praveen Talari <praveen.talari@oss.qualcomm.com>

Re: [PATCH] i2c: qcom-geni: Simplify PM resume error handling

Posted by Praveen Talari 1 month, 3 weeks ago

Hi Mukesh

On 05-08-2026 12:30, Mukesh Kumar Savaliya wrote:
> The driver currently logs pm_runtime_resume_and_get() failures with
> dev_err() and returns the error code separately.
>
> Replace this with dev_err_probe(), which combines the dev_err() log
> message and the return value propagation into a single call, consistent
> with how all other error paths in this driver already report failures.
>
> No functional change intended.
>
> Signed-off-by: Mukesh Kumar Savaliya <mukesh.savaliya@oss.qualcomm.com>
> ---
>   drivers/i2c/busses/i2c-qcom-geni.c | 6 ++----
>   1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/i2c-qcom-geni.c
> index d5c977e34994..96015e077b69 100644
> --- a/drivers/i2c/busses/i2c-qcom-geni.c
> +++ b/drivers/i2c/busses/i2c-qcom-geni.c
> @@ -994,10 +994,8 @@ static int geni_i2c_init(struct geni_i2c_dev *gi2c)
>   	int ret;
>   
>   	ret = pm_runtime_resume_and_get(gi2c->se.dev);
> -	if (ret < 0) {
> -		dev_err(gi2c->se.dev, "error turning on device :%d\n", ret);
> -		return ret;
> -	}
> +	if (ret < 0)
> +		return dev_err_probe(gi2c->se.dev, ret, "error turning on device\n");

This change is not required, as this API will be used in the deep sleep 
use case to reinitialise the SE.


Thanks,

Praveen Talari

>   
>   	proto = geni_se_read_proto(&gi2c->se);
>   	if (proto == GENI_SE_INVALID_PROTO) {

Re: [PATCH] i2c: qcom-geni: Simplify PM resume error handling

Posted by Mukesh Savaliya 1 month, 3 weeks ago

Thanks Praveen for the review !

On 8/5/2026 3:07 PM, Praveen Talari wrote:
> Hi Mukesh
> 
> On 05-08-2026 12:30, Mukesh Kumar Savaliya wrote:
>> The driver currently logs pm_runtime_resume_and_get() failures with
>> dev_err() and returns the error code separately.
>>
>> Replace this with dev_err_probe(), which combines the dev_err() log
>> message and the return value propagation into a single call, consistent
>> with how all other error paths in this driver already report failures.
>>
>> No functional change intended.
>>
>> Signed-off-by: Mukesh Kumar Savaliya <mukesh.savaliya@oss.qualcomm.com>
>> ---
>>   drivers/i2c/busses/i2c-qcom-geni.c | 6 ++----
>>   1 file changed, 2 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/ 
>> i2c-qcom-geni.c
>> index d5c977e34994..96015e077b69 100644
>> --- a/drivers/i2c/busses/i2c-qcom-geni.c
>> +++ b/drivers/i2c/busses/i2c-qcom-geni.c
>> @@ -994,10 +994,8 @@ static int geni_i2c_init(struct geni_i2c_dev *gi2c)
>>       int ret;
>>       ret = pm_runtime_resume_and_get(gi2c->se.dev);
>> -    if (ret < 0) {
>> -        dev_err(gi2c->se.dev, "error turning on device :%d\n", ret);
>> -        return ret;
>> -    }
>> +    if (ret < 0)
>> +        return dev_err_probe(gi2c->se.dev, ret, "error turning on 
>> device\n");
> 
> This change is not required, as this API will be used in the deep sleep 
> use case to reinitialise the SE.
> 

I think this change is generic and does not appear to introduce any new 
functionality. Its more like a cleanup or modernization that aligns the 
PM handling with the pattern already used in the drivers probe path.

What exactly is getting broken during deep sleep with this change ? 
Could you elaborate on the specific suspend/resume issue that is getting 
disturbed ?

> 
>>       proto = geni_se_read_proto(&gi2c->se);
>>       if (proto == GENI_SE_INVALID_PROTO) {