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.

TMS320DM368: Resource allocation problem when using Dei

Part Number: TMS320DM368

Hello All,

I am using a DVSDK version 4_02_00_01,  Framework components version 2_26_00_01,  Codec Engine version 2_26_02_11 and DMAI version 2_20_00_15.

I need to deinterlace incoming interlaced i/p from camera. For this i have used dei from RDK since there was no support for dei in DVSDK 4_02_00_01.

When i am trying to run my application there is an issue with allocation of resources. I suppose this is coming from IRES getResourceDesciptor or initResources.

I am using h264 and jpeg encoding with group id 1. Dei with group id 2.

Is this a problem with EDMA resource allocation or memory allocation or any other shared resource?

Is there a way by which I can find exactly which shared resource is not being allocated?

Thanks and Regards

  • Hi,

    What is the IPNC RDK version you are using? Please share the error logs.

    Regards,
    Shwetha
  • IPNC RDK Version 5.1.0
  • Can i get a flow chart of how HDVICP Sync (Framework Component) is working?
    Is there any problem in creating DEI resources prior to VENC1 resources?
    Group id of h264 is 1 and dei is 2.
    Please if you could reply asap. That would be very helpful
  • Hi,

    h264_Encoder and DEI share few resources. 

    Uma Shanker Gupta said:
    I am using h264 and jpeg encoding with group id 1. Dei with group id 2.
    So can you try everything under same group id?

  • My application flow is

    Capture buffer from Video Input -> Deinterlace -> Resize -> Encode -> and write the encoded buffer.

    by reading  DM36x_DEI_ReleaseNotes.pdf (DEI Algorithm on DM36x Version 2.00.00 Release Notes October 2012) i understand that process call to Deinterlacing algorithm is asynchronous and may update the output buffer when the sequencer receives interrupt from algorithm.

    Input buffer passed to the Deinterlacing algorithm is shared between some other process also and the output buffer from DEI is used by resizer.

    Since DEI process is asynchronous passing Output buffer to resizer will cause a problem?

    Or Input buffer shared between threads can cause a problem?

    Application comes in deadlock condition some times or gives a segmentation fault. I am assuming this is caused due to asynchronous processing of DEI.

    Kindly help me with the issue.