SK-TDA4VM: Unable to use more than 2 UDMA channels for C7x TDA4VM board

Hi,

the error is "resource allocation failure", so there is not sufficient resource allocated for this application to run. Can you please below FAQ and allocate enough channels on this core and then try running this example? 

(99+) [FAQ] TDA4VM: Workflow and Resource allocation build flow for SysConfig Tool? - Processors forum - Processors - TI E2E support forums

Regards,

Brijesh

Hi 

I followed the link which you shared and installed the sys config tool for windows , I couldn't find the UDMA option from the peripheral settings to change as per the suggestions given in the reply . 

bcc : d.jenkins@ti.com 

Do you have any other options to perform sys config tool configurations for UDMA channel settings

Thanks,

Madhu

Hi Madhu,

No, unfortunately, we need to allocate enough resources in RM and then only this updated example will work. 

There may not be exactly UDMA channels, but can you search and see if there is any DMA TX/RX channels that can be taken from some other core and allocated to C7x? 

Regards,

Brijesh

Hi 
Thanks for you reply 
In c66x dsp core the  block copy channels number is set as 8 from sciclient_defaultBoardCfg_rm because of which I am able to use 8 UDMA channels and for C7x block copy channels number is set as only 2 from sciclient_defaultBoardCfg_rm which gives 2 UDMA channels only .

Can we deduct 6 block copy channels number from C66x dsp and add it to the C7x dsp ? If so What is the exact procedure to do so using sys config tool?

Thanks,

Madhu

Hi Madhu,

Yes, you can take block copy channels from C6x and allocate them to C7x. 

We thought that C7x will just use DRU channels, so allocated less number of Block copy channels to C7x. 

Regards,

Brijesh

Hi 

There is a file named  sciclient_defaultBoardCfg_rm.c where num_resource have been allocated for TISCI_HOST_ID_C7X_1 and TISCI_HOST_ID_C6X_0_1, can we change the values here and check for the UDMA channels or how to set the values using sys config tool?

Regards,

Madhu