This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Running OMTB second time: No component named "OMX.TI.DUCATI.VIDDEC" found

Guru 10685 points

I am playing around with the OMTB.

First I bring down the graphics planes:

echo 0 > /sys/devices/platform/vpss/graphics0/enabled
echo 0 > /sys/devices/platform/vpss/graphics1/enabled
echo 0 > /sys/devices/platform/vpss/graphics2/enabled

Then I run the OMTB a first time with the "decode_scale_display.oms" script. This runs fine.

I then need to exit the script otherwise it runs forever so I use Ctrl-C.

I then try to run OMTB with the same script again but it errors out with:

Wait for VM3 to to create the server first
Assertion at Line no: 420 in /swcoe/sdk/cm/netra/arago-tmp/work/dm816x-evm-none-linux-gnueabi/ti-syslink-02_00_00_68-r3i/syslink_02_00_00_68_beta1/ti/syslink/utils/hlos/knl
/Linux/../../../../../../ti/syslink/ipc/hlos/knl/Linux/MessageQDrv.c: (cargs.args.create.handle != NULL) : failed
OMTB> Processing : omx -s /temp.oms
    
OMTB> omx api init
Wait for VM3 to to create the server first
Assertion at Line no: 420 in /swcoe/sdk/cm/netra/arago-tmp/work/dm816x-evm-none-linux-gnueabi/ti-syslink-02_00_00_68-r3i/syslink_02_00_00_68_beta1/ti/syslink/utils/hlos/knl
/Linux/../../../../../../ti/syslink/ipc/hlos/knl/Linux/MessageQDrv.c: (cargs.args.create.handle != NULL) : failed
OMTB-FAIL: OMX_ErrorUndefined
OK
    
OMTB> omx setp 0 h264vdec frame_width 1920
OK

* * * * * lots of text from the script which says "OK" * * * * *

OMTB> omx setp 0 dc insmod_hdmi disable
OK
    
OMTB> omx api gethandle h264vdec 0 0
OMTB-No component named "OMX.TI.DUCATI.VIDDEC" found
OMTB-FAIL: OMX_ErrorNotReady
OMTB-Setting all port params
Segmentation fault

Anyone else seen this problem? I don't want to have to reboot the board every time I run a new script.

Any help much appreciated.

Ralph

 

  • Hi Ralph, I think you have to power cycle the board in order to run the script the second time. This is a limitation in current sdk release.

  • Hi Perry,

    thanks for the reply. I'll hang on for an official TI response but in the meantime your response is a comfort.

    Regards,
    Ralph

  • Just had a thought: is this problem anything to do with what it says in the software developer's guide where it says:

    "Note: Due to the limitation with OMX_deinit() and OMX_FreeHandle() not being supported, if OMTB was
    previously run then the board will require a power cycle before OMTB can be run manually."

    ?

    I have an additional query as well as the one I first posted in this thread: how can I run OMTB without running a script, i.e. in interactive mode?

    Thanks,
    Ralph

  • yes, that's right. Application does not exit due to issues you mentioned.

    For OMTB, it requires some changes in omtbmain.c

    strcpy(inputCmdLine, "omx -s ");
          strcat(inputCmdLine, script_file);


    needs to be removed and gets(inputCmdLine); is required to be enabled to take the commands in inetractive manner.

  • Hi Vimal,

    thanks for replying. I'm glad you have confirmed what I thought about the application not exiting properly.

    As for the interactive changes, I believe that the while loop that loops over inputCmdLine would need to be found and the "gets(inputCmdLine);" added there to take a new line from user input, one line at a time. While I could implement this myself given some time, it would be nice if TI would implement this themselves in the next version of the EZSDK.

    Thanks,

    Ralph

  • Hi Ralph,

    We will enable this in next SDK release. In addition to that, you will have more eamples (without OMTB) as well.