Re: [PATCH] dmaengine: idxd: Fix memory leak in idxd_alloc()

[Dave Jiang <dave.jiang@intel.com>]

On 9/16/2022 8:36 AM, Jerry Snitselaar wrote:
> On Wed, Sep 14, 2022 at 08:08:14PM -0300, Rafael Mendonca wrote:
>> If the IDA id allocation fails, then the allocated memory for the
>> idxd_device struct doesn't get freed before returning NULL, which leads to
>> a memleak.
>>
>> Fixes: 47c16ac27d4c ("dmaengine: idxd: fix idxd conf_dev 'struct device' lifetime")
>> Signed-off-by: Rafael Mendonca <rafaelmendsr@gmail.com>
> I think there needs to be a kfree(idxd) where it checks rc < 0 after the call to dev_set_name() as well, yes?
The idxd_conf_device_release() should take care of freeing idxd with the 
put_device(). So I think we are good here.
>
> Regards,
> Jerry
>
>> ---
>>   drivers/dma/idxd/init.c | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/dma/idxd/init.c b/drivers/dma/idxd/init.c
>> index aa3478257ddb..fdc97519b8fb 100644
>> --- a/drivers/dma/idxd/init.c
>> +++ b/drivers/dma/idxd/init.c
>> @@ -445,8 +445,10 @@ static struct idxd_device *idxd_alloc(struct pci_dev *pdev, struct idxd_driver_d
>>   	idxd->data = data;
>>   	idxd_dev_set_type(&idxd->idxd_dev, idxd->data->type);
>>   	idxd->id = ida_alloc(&idxd_ida, GFP_KERNEL);
>> -	if (idxd->id < 0)
>> +	if (idxd->id < 0) {
>> +		kfree(idxd);
>>   		return NULL;
>> +	}
>>   
>>   	device_initialize(conf_dev);
>>   	conf_dev->parent = dev;
>> -- 
>> 2.34.1
>>