TDA4VM: UDMA 2D setup: Should dicnt0 always be equal to icnt0?

Hi FredC_LT,

Are you using any trigger or Event functionality on icnt0? Then in this case, both src and dst parameters must be same. 

Regards,

Brijesh

i Brijesh,

I'm using the vision_apps app_udma API.

Specifically:

• appUdmaCopyNDInit()
• appUdmaCopyNDTrigger()
• appUdmaCopyNDWait()

appUdmaCopyNDInit() initializes a CSL_UdmapTR15* with the following flags:

pTr->flags  = CSL_FMK(UDMAP_TR_FLAGS_TYPE, CSL_UDMAP_TR_FLAGS_TYPE_4D_BLOCK_MOVE);
pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_STATIC, 0U);
pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_EOL, 0U);

pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_EVENT_SIZE, CSL_UDMAP_TR_FLAGS_EVENT_SIZE_COMPLETION);
pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_TRIGGER0, CSL_UDMAP_TR_FLAGS_TRIGGER_NONE);
pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_TRIGGER0_TYPE, CSL_UDMAP_TR_FLAGS_TRIGGER_TYPE_ALL);


pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_TRIGGER1, CSL_UDMAP_TR_FLAGS_TRIGGER_NONE);
pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_TRIGGER1_TYPE, CSL_UDMAP_TR_FLAGS_TRIGGER_TYPE_ALL);

pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_CMD_ID, 0x25U);
pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_SA_INDIRECT, 0U);
pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_DA_INDIRECT, 0U);
pTr->flags |= CSL_FMK(UDMAP_TR_FLAGS_EOP, 1U);

Hi FredC_LT,

ok, this does not enable any trigger or event, so it should work. Let me check with team if there is any other restriction.

Regards,

Brijesh

Hi FredC_LT,

Can you please try with below highlighted change? 

int32_t Col2Row_4x4(float* src, float* dst)

{

    app_udma_copy_nd_prms_t prms_nd;

    int32_t ret = 0;

    /* Set Up the Copy */

    appUdmaCopyNDPrms_Init(&prms_nd);

    prms_nd.copy_mode = 0;

    prms_nd.eltype = 0;

    /* Source Params */

    prms_nd.icnt0 = sizeof(float);

    prms_nd.icnt1 = 4;

    prms_nd.icnt2 = 1;

    prms_nd.icnt3 = 1;

    prms_nd.dim1 = sizeof(float) * 4; // size of element * number of elements in row.

    prms_nd.dim2 = 0;

    prms_nd.dim3 = 0;

    prms_nd.src_addr = (uint64_t) src;

    /* Destination Params */

    prms_nd.dicnt0 = sizeof(float) * 4;

    prms_nd.dicnt1 = 1;

    prms_nd.dicnt2 = 1;

    prms_nd.dicnt3 = 1;

    prms_nd.ddim1 = 0;

    prms_nd.ddim2 = 0;

    prms_nd.ddim3 = 0;

    prms_nd.dest_addr = (uint64_t) dst;

    ret = appUdmaCopyNDInit(udmaCpyCh, prms_nd);

    if (0 == ret)

    {

        ret = appUdmaCopyNDTrigger(udmaCpyCh);

        if (0 == ret)

        {

            ret = appUdmaCopyNDWait(udmaCpyCh);

        }

        appUdmaCopyNDDeinit(udmaCpyCh);

    }

    return ret;

}

That was probably initially a typo. I double-checked and it is what I'm using.

    prm.icnt0 = sizeof(float);
    prm.icnt1 = 4;
    prm.icnt2 = 1;
    prm.icnt3 = 1;
    prm.dim1 = sizeof(float) * 4;
    prm.dim2 = 0;
    prm.dim3 = 0;

    // set the dest parameters
    prm.dicnt0 = sizeof(float) * 4;
    prm.dicnt1 = 1;
    prm.dicnt2 = 1;
    prm.dicnt3 = 1;
    prm.ddim1 = 0;
    prm.ddim2 = 0;
    prm.ddim3 = 0;

Output:

src: {
  { 1, 2, 3, 4 },
  { 1, 2, 3, 4 },
  { 1, 2, 3, 4 },
  { 1, 2, 3, 4 },
}
dst: {
  { 1, 0, 0, 0 },
  { 0, 0, 0, 0 },
  { 0, 0, 0, 0 },
  { 0, 0, 0, 0 },
}

If I use this instead:

    prm.icnt0 = sizeof(float);
    prm.icnt1 = 4;
    prm.icnt2 = 1;
    prm.icnt3 = 1;
    prm.dim1 = sizeof(float) * 4;
    prm.dim2 = 0;
    prm.dim3 = 0;

    // set the dest parameters
    prm.dicnt0 = sizeof(float);
    prm.dicnt1 = 4;
    prm.dicnt2 = 1;
    prm.dicnt3 = 1;
    prm.ddim1 = sizeof(float);
    prm.ddim2 = 0;
    prm.ddim3 = 0;

Then it works:

src: {
  { 1, 2, 3, 4 },
  { 1, 2, 3, 4 },
  { 1, 2, 3, 4 },
  { 1, 2, 3, 4 },
}
dst: {
  { 1, 1, 1, 1 },
  { 0, 0, 0, 0 },
  { 0, 0, 0, 0 },
  { 0, 0, 0, 0 },
}

Hi Brijesh, did you have a chance to review?

Hi Fred,

Brijesh will review and follow-up on this in the next couple of days.

regards

Suman