I wrote, while trying to compile the wb_dma testbench, about
VCSi errors of the form...

The problem is that there is a 'wb_mast m0(...)' definition in bench/verilog/test_bench_top.v, but in the `included tests.v, there is at least one definition of an *integer* m0. Yurk! Different simulators do different things. 'cver' gives the best error messages; VCSi is right to spot a problem but gives a terrible error message; both icarus and an old verilog-xl just stagger forward blindly (and the simulations are completely useless). Anyway, the attached patch fixes the problem. At the very end of the patch, there's an ifndef-thing around the contents of rtl/verilog/wb_dma_defines.v -- the sort of thing that is standard in C programming. It eliminates some compiler warnings, and I commend it to you. Keep up the good work, Will PS: Shouldn't there be one or more 'bug reporting' links/pages/whatever on www.opencores.org? -------------- next part -------------- A non-text attachment was scrubbed... Name: wb_dma.diffs Type: application/octet-stream Size: 24130 bytes Desc: fix m0 bugs + one ifndef Url : http://www.opencores.org/forums.cgi/cores/attachments/20040206/e00386d1/wb_dma.obj

The process is much simpler. You send bug report to the author.

regards,
Damjan




----------------------------------------------------------------------------
----