Forwarding to the mailing list so other people can comment.
Thank you for your interest in the Navre core!
S.

--- Begin Message ---
Hi, I list some suggestion about navre mcu:

1. A bug, when unsupported instruction runs, GPR reg may be written. May can
changed as this :
        436,440c442,446
        <                               16'b0000_0000_0000_0000: begin
        <                                       /* NOP */
        <                                       update_nsz = 1'b0;
        <                                       writeback = 1'b0;
        <                               end
        ---
        >                               //16'b0000_0000_0000_0000: begin
        >                               //      /* NOP */
        >                               //      update_nsz = 1'b0;
        >                               //      writeback = 1'b0;
        >                               //end
        467a474,477
        >                               default: begin /* NOP and other*/
        >                                       writeback = 1'b0;
        >                                       update_nsz = 1'b0;
        >                               end

2. For the lower 96(or 256) data memory locations the GPR register file and
IO register file, so the access the lower address should map to GPR or IOF.
3. I have optimized Multi-cycle operation sequencer, please see the
attachment.
4. In Asic design, Not recommended using X(such as R16 = 16'hxxxx), for
mismatching in pre/post simulation.
5. Suggestion, The line242, suggest split 2 blocks, one is  Nonblocking
assignment and anther is blocking assignment.
6. Suggestion, add a navre enable signal for used as gate signal in asic
synthesis(such as: compile -gated).

Thanks

Attachment: softusb_navre.v
Description: Binary data


--- End Message ---
_______________________________________________
http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org
IRC: #milkym...@freenode
Twitter: www.twitter.com/milkymistvj
Ideas? http://milkymist.uservoice.com

Reply via email to