Open zhenhaohe opened 2 months ago
Hi @zhenhaohe or @quetric, can we maybe get a bit more context on this issue? Which scenario, which busses are we talking about? Do you maybe even have a screenshot of waveforms or something like this?
This bug is observed on the cq_wr
bus. I don't have a screenshot of waveforms, but the context is that after a write request is issued on sq_wr
, we are observing cq_wr
for a completion. The bug is that instead of TVALID staying high for 1 cycle to signal a completion, it stays high for 2 consecutive cycles.
Hi @quetric, thanks a lot for the clarification. We'll be looking into it.
Hi all,
Recently I found a potential bug with the completion queue of Coyote. Each compeletion of the request comes with a 2-cycle burst valid signal instead of a 1-cycle valid.