Conversation
aolofsson
left a comment
aolofsson
left a comment
There was a problem hiding this comment.
Looks good. @RiceShelley might have some comments on cocotb structure.
In terms of PR, better to split out PRs into smaller ones. It would have been better to have the verilog in separate smaller PR so that I could merge that immediately.
RiceShelley
left a comment
RiceShelley
left a comment
There was a problem hiding this comment.
A few changes requested. If it helps unblock things, we can spin a second PR with just the RTL to get that checked in earlier.
Great job jumping in and learning cocotb in just one week, @AndyOja !
| ); | ||
|
|
||
| // UMI memory agent (handles UMI requests and generates responses) | ||
| umi_memagent #( |
There was a problem hiding this comment.
It would be good to test the UMI side of tl2umi against something other than umi_memagent. The ready/valid behavior of umi_memagent likely doesn’t exercise the full space of possible handshake timings.
For example what if we wanted to test that when back pressure is applied to umi_req it propagates back to the tl_a interface correctly?
In the future I would like to see a test of tl2umi where the UMI interface is driven from cocotb with constrained random ready / valid signal timing.
AndyOja
force-pushed
the
andy/adapters_tests_updates
branch
from
39249d8 to
a3730b5
Compare
3 participants
Added cocotb testing for umi2apb and tl2umi and debugged umi2apb umi response and tl2umi FSM.