OpenCores

Improve test bench around Memory functionality

Back to bugtracker overview.

Information:
Type :: REQUEST
Status :: CLOSED
Assigned to ::

Description:
Just 1 exemple,

File : dmem_ctrl.vhd v1.5

153 -- Process dmem_addr_reg
154 --
155 -- Purpose:
156 -- Implements the Data Memory Address Register.
157 -- This register is required to hold the address during a write operation
158 -- as we cannot hold the address in the input register of the
159 -- synchronous RAM (no clock suppression/gating).
160 --
161 -- NOTE: May be obsoleted by clock enable feature of generic RTL RAM.
162 --
163 dmem_addr_reg: process (res_i, clk_i)
164 begin
165 if res_i = res_active_c then
166 dmem_addr_q <= (others => '0');

If I stick always the reset condition (l.165) to true, no test bench detect it.

Can you improve the test env for take care of memory functionality ?
Because there is lot of verification hole around memory features like on dmem_addr, dmem_o or dmem_read ...

Thanks a lot for all your work, and good luck.

Julien

Comments:

Laeuger, Arnim Feb 23, 2008
I can't reproduce that a constant reset condition in line 165 passes all tests. Some regression tests fail (e.g. black_box/mov/a_rr/data_00), some timeout (e.g. black_box/xch/rr) when the condition is changed to
165 if true then

Wrong functionality on dmem_addr_q is detected in my opinion, or did I miss something in your post?
Laeuger, Arnim Feb 3, 2008
Thanks for reporting. I'll have a look at this topic within the next days.

Post a comment:
Login to post comments!

Back to bugtracker overview.

© copyright 1999-2012 OpenCores.org, equivalent to ORSoC AB, all rights reserved. OpenCores®, registered trademark.