On 10/28/2016 10:34 AM, CK Hu wrote: > Hi, Matthias: > > Even though OVL HW would not be enabled before component_add() in > current design, your patch would be safe for any situation.
Maybe the FW I use left an interrupt pending before loading the kernel and this leads to the case where we enter the handler before all data structures are set up. > > Acked-by CK Hu <ck.hu at mediatek.com> Thanks! Matthias
