How to handle on-the fly reset scenario in verification component

OVM
,
1

i All,

I would like to know how to handle reset in OVM based verification component.

Please find my scenario:

  1. My component will be working as Slave-responder model, where all transactions initiated by DUT
  2. Responder will give back the response based on the command
  3. I have a test case where reset will be applied randomly in middle of data traffic
  4. I am getting below error from sequencer:
    Get_next_item called twice without item_done or get in between
  5. Looks from driver we got item to drive, before its getting drive system got a reset and is applied to our agent as well, so driver was not able to give back item_done
  6. Post to reset I am trying to initiate new traffic, where I am getting the above error.

Can some one help me how to handle or implement logic to take of on-the fly reset scenario.
It would be great if you point me to some reference code.

Thanks,
Deepak

2

In reply to deepak_infy:

hi
as per your requirement i understand the following , your dut is putting a req, and your tb response back , lets say dut_req is request from dut , and tb_resp is reponse to dut

now


//initial code
forever begin
seq_item_port.get_next_item(req);
drive_item();
seq_item_port.item_done(); // this will be run even after reset as, after drive_item() task overs due to reset assertion
end

//now inside drive_item put like this
task drive_item();
fork
begin: driver_code
real_driver_job();
real_driver_job2();
     wait(vif.dut_req); // this is a point where you are waiting for the DUT req and reset may happen.
end: driver_code
begin: reset_handler
@ negedge vif.reset ; 
//give uvm_info for reset assertion
disable driver_code ; // this disable the current transaction of the driver
//now drive the reset value to the DUT , ie drive the reset response
vif.tb_resp = 0 ;
end: reset_handler
join_any

endtask