[PATCH] media: s5p-mfc fix memory leak in s5p_mfc_remove()

Shuah Khan shuahkh at osg.samsung.com
Mon Jun 13 08:38:38 PDT 2016


On 06/13/2016 07:42 AM, Javier Martinez Canillas wrote:
> Hello Shuah,
> 
> On Wed, Jun 8, 2016 at 9:35 PM, Shuah Khan <shuahkh at osg.samsung.com> wrote:
>> s5p_mfc_remove() fails to release encoder and decoder video devices.
>>
>> Signed-off-by: Shuah Khan <shuahkh at osg.samsung.com>
>> ---
>>  drivers/media/platform/s5p-mfc/s5p_mfc.c | 2 ++
>>  1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/media/platform/s5p-mfc/s5p_mfc.c b/drivers/media/platform/s5p-mfc/s5p_mfc.c
>> index 274b4f1..af61f54 100644
>> --- a/drivers/media/platform/s5p-mfc/s5p_mfc.c
>> +++ b/drivers/media/platform/s5p-mfc/s5p_mfc.c
>> @@ -1317,7 +1317,9 @@ static int s5p_mfc_remove(struct platform_device *pdev)
>>         destroy_workqueue(dev->watchdog_workqueue);
>>
>>         video_unregister_device(dev->vfd_enc);
>> +       video_device_release(dev->vfd_enc);
>>         video_unregister_device(dev->vfd_dec);
>> +       video_device_release(dev->vfd_dec);
>>         v4l2_device_unregister(&dev->v4l2_dev);
>>         s5p_mfc_release_firmware(dev);
>>         vb2_dma_contig_cleanup_ctx(dev->alloc_ctx[0]);
>> --
> 
> Can you please do the remove operations in the inverse order of their
> counterparts? IOW to do the release for both encoder and decoder after
> their unregistration.
> 

I considered that. No problem. I will move both release after the
video_unregister_device(dev->vfd_dec), releasing enc first and then
the dec

> 
> Reviewed-by: Javier Martinez Canillas <javier at osg.samsung.com>
> 

thanks,
-- Shuah




More information about the linux-arm-kernel mailing list