VPIF RAW mode driver reference code ofr DM6467

We have a couple of docs that are included with the DVSDK installation (look under PSP_XX_XX_XXX/docs directory) that describe VPIF capture and display driver features correspondingly (SPRUG99 and SPRUGA0)

I want to use VPIF RAW capture mode of DM6467. Actually I send data/clock from FPGA with R/G/B raw data format(10-bits) and want to use VPIF RAW capture mode to get these datas, can it work? If yes, how will I begin with raw capture mode? Is there any example to refer?

Mark Chang said:

I want to use VPIF RAW capture mode of DM6467. Actually I send data/clock from FPGA with R/G/B raw data format(10-bits) and want to use VPIF RAW capture mode to get these datas, can it work? If yes, how will I begin with raw capture mode? Is there any example to refer?

Hi Mark

             Did you find any solution to your query even iam at same situation and don't know how to move ahead.

Keep in mind you will need Hblank and Vblank signal. Other than that, raw capture is very flexible as it just takes in whatever you feed it and stores them in  the memory buffer. However, it won't be any processing before storing.

Were you looking for simple C initialization code or Linux example?

hi Paul

           Like Mark iam also in same situation where i dont have a decoder in direct access of Davinci and is passing data through a FPGA interfaced to VPIF. So if i could have some example for Linux to look into, it will be of great help .

Hi,

I am not sure what you meant by "i dont have a decoder in direct access of Davinci and is passing data through a FPGA interfaced to VPIF" Are you saying your raw data comes from FPGA instead of a A/D?

I am not able to find an actual raw capture demo, I'd suggest you to read SPRUG99.pdf like Juan suggested above.

Hello,

We have here a very similar problem:

We are using the DM6467 chip on a daughter board that does not have TVP514X decoder nor CPLD. Instead, the DSP receieves BT656 data directly to VPIF. When we launch our application we get an error in the VPIF driver because it tries to configure the TVP514X that does not exist. We modified the file tvp514x.c in the linux kernel to solve the problem by bypassing any attempt to access the I2C to configure the "ghost" TVP514X. But nevertheless, we still get this error:

VIDIOC_S_INPUT

: Remote I/O error

Error in opening capture device for channel 0

We have discovered that this error comes from a (yet another) access to the TVP514X through I2C (in i2c-davinci.c)

At this point we wonder:

1.- Is there a way to make the configuration of the VPIF port independent of any access to I2C?

2.- Are we in the good direction? I mean: is modifying the linux drivers the best approach to solve this problem?

3.- Is there an easier way to configure the VPIF port for our purposes, other than tweaking the drivers?

4.- We read here about the RAW capture mode, can it be useful in our case? How?

Thanks,

César

Hi Cesar,

This is a separate Linux related issue, please open a new thread in the Linux forum.

Ok, I re-posted my question here:

http://e2e.ti.com/support/embedded/f/354/p/70989/257484.aspx#257484