TLM Channels in OVM :

1. tlm_fifo

unidirectional transactions between free-running independent components

// Class 
class tlm_fifo #(type T, int BOUND=1);
    local mailbox #(T) m;
    extern task put(input T t);
    extern task get(output T t);
    extern task peek(output T t);
    ...
endclass

2. tlm_req_rsp_channel

Two back-to-back tlm_fifos pipelined or out-of-order protocols

class tlm_req_rsp_channel

    #(type REQ, RSP, int BOUND=1);

    tlm_fifo #(REQ, BOUND) req = new();

    tlm_fifo #(RSP, BOUND) rsp = new();
    ...
endclass

3. tlm_transport_channel

1. tlm_req_rsp_channel with fifo size = 1
2. non-pipelined and in-order

class tlm_transport_channel     #(type REQ, RSP);

    tlm_fifo #(REQ, 1) req = new();
    tlm_fifo #(RSP, 1) rsp = new();
    ...

    task transport(input REQ reqt, output RSP rspt);
      req.put(reqt);
      rsp.get(rspt);
      ...
    endtask 

endclass

Analysis Fifo in SystemVerilog

Specialization of tlm_fifo
Infinite fifo to ensure nonblocking writes
Still has tlm_*_get exports on the other side

class analysis_fifo #( type T = int ) extends tlm_fifo #( T );
    analysis_imp #( analysis_fifo #( T ) , T ) analysis_export;
    function new( string nm , named_component p = null );
        super.new( nm , p , 0 );
        analysis_export = new( this );
    endfunction
    function void write( input T t );
        try_put( t );
    endfunction
endclass

This Articles is written/submitted by puneet (Puneet Aggarwal). You can also contribute to Asicguru.com. Click here to start