I'm testing a linux + codec-engine based system on our 6446 board. Lately, it seems that the algorithm will stall after running a few hours. I turned the CE_DEBUG=2 on, then it showed where the program stalled as below:
...
[DSP] @3829,345,939tk: [+5 T:0x878dae4c] CV - VISA_exit(visa=0x878dadd8): algHandle = 0x878dae08
[DSP] @3829,346,020tk: [+0 T:0x878dae4c] ti.sdo.ce.alg.Algorithm - Algorithm_deactivate> Enter(alg=0x878dae08)
[DSP] @3829,346,098tk: [+0 T:0x878dae4c] ti.sdo.ce.alg.Algorithm - Algorithm_deactivate> Exit
[DSP] @3829,346,159tk: [+0 T:0x878dae4c] ti.sdo.ce.video1.VIDENC1 - VIDENC1_process> Exit (handle=0x878dadd8, retVal=0x0)
[DSP] @3829,346,245tk: [+0 T:0x878dae4c] OM - Memory_cacheWb> Enter(addr=0x859d4000, sizeInBytes=129288)
[DSP] @3829,346,597tk: [+0 T:0x878dae4c] OM - Memory_cacheWb> return
[DSP] @3829,346,646tk: [+5 T:0x878dae4c] CN - NODE> returned from call(algHandle=0x878dadd8, msg=0x87d06d00); messageId=0x00020c01
@1058,084,171us: [+0 T:0x439a8b60] CE - Engine_fwriteTrace> returning count [2138]
@1058,084,469us: [+0 T:0x439a8b60] CV - VISA_call Completed: messageId=0x00020c01, command=0x0, return(status=0)
@1058,084,794us: [+5 T:0x439a8b60] CV - VISA_freeMsg(0x53fd8, 0x414afd00): Freeing message with messageId=0x00020c01
@1058,085,106us: [+0 T:0x439a8b60] ti.sdo.ce.video1.VIDENC1 - VIDENC1_process> Exit (handle=0x53fd8, retVal=0x0)
@1058,164,077us: [+0 T:0x439a8b60] ti.sdo.ce.video1.VIDENC1 - VIDENC1_process> Enter (handle=0x53fd8, inBufs=0x439a8420, outBufs=0x439a8410, inArgs=0x439a8400, outArgs=0x439a8388)
@1058,164,591us: [+5 T:0x439a8b60] CV - VISA_allocMsg> Allocating message for messageId=0x00020c02
@1058,164,936us: [+0 T:0x439a8b60] CV - VISA_call(visa=0x53fd8, msg=0x414afd00): messageId=0x00020c02, command=0x0
Since the last printed log shows the program entered VISA_call. I assum that it had something to do with messege passing.
Am i guessing it right? Anyone point me a direction to debug this, please.