PDA

View Full Version : IN case of OMX_CONF_WaitForResourcesTest



jaeryul
08-21-2008, 10:39 PM
I'm testing conformance of openMAXIL. but in case of WaitForResourceTest,
Error happens in the line#453(OMX_CONF_WaitForResourcesTest.c) because of
the more instantiation(my MAX_COMPONENT_VIDEODEC = 4), then

My question is that next seq. (4~5 as below) is not progressed.
/* ---------------------------------------
* 1. Load component
* 2. Transition to idle
* 3. Repeat from steps 1-3 until transition to idle returns
* OMX_ErrorInsufficientResources and transitions to Wait state
* 4. Switch one of the previous components to the loaded state to free resources
* 5. Verify the component in the WaitForResources state transitions to idle
*/

My log is as below, What is the problem ? & What should I check ?

------------------------------------------------------------
OMX_CONF_AddTest WaitForResourcesTest
OMX_CONF_TestComponent OMX.sec.video_decoder.mpeg4
##
## WaitForResourcesTest
##
-Trying to create and idle instantiation 0 (priority = 100)
-Instantiation 0 changing to state OMX_StateIdle
-Trying to create and idle instantiation 1 (priority = 100)
-Instantiation 1 changing to state OMX_StateIdle
-Trying to create and idle instantiation 2 (priority = 100)
-Instantiation 2 changing to state OMX_StateIdle
-Trying to create and idle instantiation 3 (priority = 100)
-Instantiation 3 changing to state OMX_StateIdle
-Trying to create and idle instantiation 4 (priority = 100)
-!!! InteropProfileTests/OMX_CONF_WaitForResourcesTest.c:453 Error 80001000 OMX_ErrorInsufficientResources
-Cleaning up
-Instantiation 1 changing to state OMX_StateLoaded
-Instantiation 2 changing to state OMX_StateLoaded
-Instantiation 3 changing to state OMX_StateLoaded
*** WaitForResourcesTest OMX.sec.video_decoder.mpeg4 FAILED, 80001000 OMX_ErrorInsufficientResources

***
*** Summary for OMX.sec.video_decoder.mpeg4
***
*** 0 tests passed:
***
*** 1 tests failed:
*** WaitForResourcesTest
***

rnixon
08-22-2008, 08:49 AM
Your line numbers don't seem to match the latest adopters' package. I suspect that you're limiting the number of instances in OMX_GetHandle() rather than failing to allocate resources on the transition to idle (all ports disabled, so no buffers, but hardware resources should be allocated). In order to perform resource preemption, you need to be able to create more instances than can be transitioned to idle, i.e. than can be accomodated by the hardware.

It's the transitioning to idle that allocates the resources, so it's here that an insufficient resources error event is expected to be raised. We then up the priority of the last (failing) instance and see that it preempts resources to suceed the transition to idle.

Thanks,

Roger