Skip to content

Fix APU tracer clock and reset connections - #1087

Open
jinchuntan wants to merge 1 commit into
openhwfoundation:devfrom
jinchuntan:fix/apu-tracer-clock-reset
Open

jinchuntan wants to merge 1 commit into
openhwfoundation:devfrom
jinchuntan:fix/apu-tracer-clock-reset

Conversation

@jinchuntan

Copy link
Copy Markdown

The APU tracer has its clock and reset connections swapped in cv32e40p_tb_wrapper. It opens its log on a clock edge and then waits for reset transitions to sample results. A normal FPU program therefore completes with no APU trace entries.

Connect the tracer to the core clock and active-low reset, matching the execution tracer.

Tested on the actual core and FPU with F latencies 0 and 2 and Zfinx latency 1. The final memory result is correct in all three runs and every trace destination and value matches. Unmodified dev produces only the trace header. FPU-disabled elaboration, the repository formatter and git diff --check also pass. The simulation retains baseline nonfatal warnings and the existing CSR array waiver.

Reproduction and before/after logs are attached below.
cv32e40p-apu-tracer-validation.zip

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant