Re: [PATCH] staging: vchiq_arm: Fix missing refcount decrement in error path for fw_node

[Krzysztof Kozlowski <krzk@kernel.org>]

On 13/10/2024 13:36, Umang Jain wrote:
> Hi Javier,
> 
> Thank you for the patch.
> 
> On 13/10/24 4:12 pm, Javier Carrasco wrote:
>> An error path was introduced without including the required call to
>> of_node_put() to decrement the node's refcount and avoid leaking memory.
>> If the call to kzalloc() for 'mgmt' fails, the probe returns without
>> decrementing the refcount.
>>
>> Use the automatic cleanup facility to fix the bug and protect the code
>> against new error paths where the call to of_node_put() might be missing
>> again.
>>
>> Cc: stable@vger.kernel.org
>> Fixes: 1c9e16b73166 ("staging: vc04_services: vchiq_arm: Split driver static and runtime data")
>> Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
>> ---
>>   drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c | 6 ++----
>>   1 file changed, 2 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c
>> index 27ceaac8f6cc..792cf3a807e1 100644
>> --- a/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c
>> +++ b/drivers/staging/vc04_services/interface/vchiq_arm/vchiq_arm.c
>> @@ -1332,7 +1332,8 @@ MODULE_DEVICE_TABLE(of, vchiq_of_match);
>>   
>>   static int vchiq_probe(struct platform_device *pdev)
>>   {
>> -	struct device_node *fw_node;
>> +	struct device_node *fw_node __free(device_node) =
>> +		of_find_compatible_node(NULL, NULL, "raspberrypi,bcm2835-firmware");
> 
> How about :
> 
> +	struct device_node *fw_node __free(device_node) = NULL;
> 
>>   	const struct vchiq_platform_info *info;
>>   	struct vchiq_drv_mgmt *mgmt;
>>   	int ret;
>> @@ -1341,8 +1342,6 @@ static int vchiq_probe(struct platform_device *pdev)
>>   	if (!info)
>>   		return -EINVAL;
>>   
>> -	fw_node = of_find_compatible_node(NULL, NULL,
>> -					  "raspberrypi,bcm2835-firmware");
> 
> And undo this (i.e. keep the of_find_compatible_node() call here

The point of using cleanup is to have constructor and destructor in one
place, not split. This is not in the spirit of cleanup. Linus also
commented on this and cleanup.h *explicitly* recommends not doing so. It
also lead to real bugs from time to time, so no, please do not insist on
such weird way.

Best regards,
Krzysztof