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.

TMS320DM816x Silicon Errata : Advisory 1.1.24

Other Parts Discussed in Thread: TVP7002

Hi,

 

I am trying to capture the 1080i video using OMTB on a DM8168.

But I was only able to capture 288 lines of frame.

 

I think this probrem is caused by the errata mentioned in Advisory 1.1.24 in SPRZ329.

But I don't now how to implement the "Workaround".

 

So please tell me the followings.

1. How can I configure the VPDMA max height.

2. How can I use SC in the capture path to crop the height?

3. How can I allocate extra buffer space in memory?

 

Best regards,

Syoichi Kondo

  •  

    I dont think this issue is linked to the Errata you have qouted. The workaround for the errata is already implemented in the Firmware binaries. Linux applications using the Capture functionality are transparent to this issue and the workaround.

    It would be helpful if you could qoute the relevant lines of OMTB scripts which are configuring the vfcc and TVP ctrl component. Are you trying this out on TI EVM? If not, what is the capture device you are using? Currently only TVP7002 is supported.

    Archith

  • Hi Archith,

    Thank you for your reply.

     

    I understood that the workaround is already implemented.


    The OMTB scripts is 8666.capture.zip .

    And the capure device I'm using is GS2971A manufactured by Genum Corporation.

    Is there any way to capture video with HDVPSS using GS2971A?

     

    Best regards,

    Syoichi Kondo

  • Syoichi KONDO said:

    And the capure device I'm using is GS2971A manufactured by Genum Corporation.

    Is there any way to capture video with HDVPSS using GS2971A?

     

     

    Decoders other than TVP7002 are not supported in the present EZSDK as they are programmed via the Firmware which is black-boxed.

    There are plans to support video decoder programming via Linux on A8 sometime in 4Q11 via a PSP release. 

     

    Archith 

     

  • Archith John Bency said:
    There are plans to support video decoder programming via Linux on A8 sometime in 4Q11 via a PSP release. 

    Will it be possible with only programming via Linux on A8 without an update of the Firmware?

     

    Does the TVP ctrl component just configure TVP7002?

    If I configure a decoder properly instead of the TVP ctrl component, could I capture video with current EZSDK?

     

    Best regard,
    Syoichi Kondo

  • Yes,

    Once you have the TVP7002 support through A8 you will be able to program TVP7002 without updating firmware.

     

    Regarding second question,

    Which decoder you are planning to support on your board and on which I2C bus. This is because currently all decoders are on I2C-1 and  that is being controlled through media controller firmware. So if you have decoder on same bus you wont be able to control it through Linux. This limitation will go away in 4Q11 once we will have all the encoder/decoder controlled through Linux.

     

    Regards,

    Hardik Shah

  • Hi Hardik,

    Thank you for your reply.

     

    The decoder on our board is on SPI bus. And I can read and write the registers of the decoder through Linux.

    Is there anything else I should configure to capture?

     

    Best regards,

    Syoichi Kondo

  • Hi,

    In that case you can use your driver to configure the decoder through Linux. Since the capture driver is configured for TVP7002, you should configure your decoder such that it throws data exactly same way as what TVP throws, BT1120 embedded sync. Once we have capture driver available on Linux, you will be able to configure Capture driver for, according to how your decoder is connected on board.

     

    Regards,

    Hardik Shah

  • Hi Hardik,

     

    If the decoder is configured to throw BT1120 embedded sync, is it not necessary to execute the TVP ctrl component of OpenMAX?

     

    Best regards,

    Syoichi Kondo

  • If you can ensure that the decode is pumping BT1120 embedded sync into the SoC's VIP port, then TVP comoponent is not required to be created/configured/executed. But please keep in mind that this has not been tested in the EZSDK.

     

    Archith

  • Archith John Bency said:
    But please keep in mind that this has not been tested in the EZSDK.

    I understood.

    Thanks a lot!

     

    Best regards,

    Syoichi Kondo

  • Syoichi KONDO said:

    But please keep in mind that this has not been tested in the EZSDK.

    I understood.

    Thanks a lot!

     

    Best regards,

    Syoichi Kondo

    [/quote]

    Hello Syoichi,

    Saw your thread and was curious as to if you'd managed to get your 8168 to receive and process

    non-TVP7002 originated data when you bypassed/ignored the OpenMax/TVP component from your

    OMTB scripts after all.

    We're planning on using an ADV7842 chip on our system and tripped over the whole M3 vs A8 control

    of the I2C-1 bus and "opaque" binary issue for the HDVPSS cores a while back. I had actually tried just

    commenting out the TVP portions of some OMTB scripts and then setting up our video decoder *before*

    we loaded the HDVPSS-core. That didn't work for us then so we've moved on to other portions of our

    project until we can get some control/visibility of the video-decoder setup when it moves out of the M3

    and into the A8.

     

    I'd just be curious to know if anyone out there has actually gotten this type of TVP-component bypass

    to work. I understand that it's not supported officially by TI and it hasn't been tested yet by them.

     

    Have a good weekend.

     

    Juan Solis