[PATCH] firmware: arm_scmi: fix missing destroy_workqueue() on error in scmi_notification_init

[PATCH] firmware: arm_scmi: fix missing destroy_workqueue() on error in scmi_notification_init

[Qinglang Miao]

Add the missing destroy_workqueue() before return from
scmi_notification_init in the error handling case when
fails to do devm_kcalloc().

Fixes: bd31b249692e ("firmware: arm_scmi: Add notification dispatch and delivery")
Signed-off-by: Qinglang Miao <miaoqinglang@huawei.com>
---
 drivers/firmware/arm_scmi/notify.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/firmware/arm_scmi/notify.c b/drivers/firmware/arm_scmi/notify.c
index 2754f9d01636..3048e57d9731 100644
--- a/drivers/firmware/arm_scmi/notify.c
+++ b/drivers/firmware/arm_scmi/notify.c
@@ -1476,8 +1476,10 @@ int scmi_notification_init(struct scmi_handle *handle)
 
 	ni->registered_protocols = devm_kcalloc(handle->dev, SCMI_MAX_PROTO,
 						sizeof(char *), GFP_KERNEL);
-	if (!ni->registered_protocols)
+	if (!ni->registered_protocols) {
+		destroy_workqueue(ni->notify_wq);
 		goto err;
+	}
 
 	mutex_init(&ni->pending_mtx);
 	hash_init(ni->pending_events_handlers);
-- 
2.23.0

Re: [PATCH] firmware: arm_scmi: fix missing destroy_workqueue() on error in scmi_notification_init

[Cristian Marussi]

On Mon, Nov 09, 2020 at 05:15:17PM +0800, Qinglang Miao wrote:
> Add the missing destroy_workqueue() before return from
> scmi_notification_init in the error handling case when
> fails to do devm_kcalloc().
> 
> Fixes: bd31b249692e ("firmware: arm_scmi: Add notification dispatch and delivery")
> Signed-off-by: Qinglang Miao <miaoqinglang@huawei.com>
> ---
>  drivers/firmware/arm_scmi/notify.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/firmware/arm_scmi/notify.c b/drivers/firmware/arm_scmi/notify.c
> index 2754f9d01636..3048e57d9731 100644
> --- a/drivers/firmware/arm_scmi/notify.c
> +++ b/drivers/firmware/arm_scmi/notify.c
> @@ -1476,8 +1476,10 @@ int scmi_notification_init(struct scmi_handle *handle)
>  
>  	ni->registered_protocols = devm_kcalloc(handle->dev, SCMI_MAX_PROTO,
>  						sizeof(char *), GFP_KERNEL);
> -	if (!ni->registered_protocols)
> +	if (!ni->registered_protocols) {
> +		destroy_workqueue(ni->notify_wq);
>  		goto err;
> +	}
>  

Good catch, looks good to me.

Even better you could move the above alloc_workqueue() block down here
so that you can avoid all together the additional destroy_workqueue() on
the above error path.

Thanks

Cristian


>  	mutex_init(&ni->pending_mtx);
>  	hash_init(ni->pending_events_handlers);
> -- 
> 2.23.0
> 
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH] firmware: arm_scmi: fix missing destroy_workqueue() on error in scmi_notification_init

[Qinglang Miao]

在 2020/11/10 1:51, Cristian Marussi 写道:
> On Mon, Nov 09, 2020 at 05:15:17PM +0800, Qinglang Miao wrote:
>> Add the missing destroy_workqueue() before return from
>> scmi_notification_init in the error handling case when
>> fails to do devm_kcalloc().
>>
>> Fixes: bd31b249692e ("firmware: arm_scmi: Add notification dispatch and delivery")
>> Signed-off-by: Qinglang Miao <miaoqinglang@huawei.com>
>> ---
>>   drivers/firmware/arm_scmi/notify.c | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/firmware/arm_scmi/notify.c b/drivers/firmware/arm_scmi/notify.c
>> index 2754f9d01636..3048e57d9731 100644
>> --- a/drivers/firmware/arm_scmi/notify.c
>> +++ b/drivers/firmware/arm_scmi/notify.c
>> @@ -1476,8 +1476,10 @@ int scmi_notification_init(struct scmi_handle *handle)
>>   
>>   	ni->registered_protocols = devm_kcalloc(handle->dev, SCMI_MAX_PROTO,
>>   						sizeof(char *), GFP_KERNEL);
>> -	if (!ni->registered_protocols)
>> +	if (!ni->registered_protocols) {
>> +		destroy_workqueue(ni->notify_wq);
>>   		goto err;
>> +	}
>>   
> 
> Good catch, looks good to me.
> 
> Even better you could move the above alloc_workqueue() block down here
> so that you can avoid all together the additional destroy_workqueue() on
> the above error path.
> 
> Thanks
> 
> Cristian
It sounds convincible to me Cristian, thanks for your advice ;D

I will send a v2 patch on this one in no time.
> 
> 
>>   	mutex_init(&ni->pending_mtx);
>>   	hash_init(ni->pending_events_handlers);
>> -- 
>> 2.23.0
>>
>>
>> _______________________________________________
>> linux-arm-kernel mailing list
>> linux-arm-kernel@lists.infradead.org
>> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
> .
>