Alex Williams

I don't think the host aborting commands needs to be reported to SW, but it might be good to upload only commands that are a valid length.

Now that we have hardware handling WREN / WRDI, we cover the common use cases with the capability to handle just one uploaded command. Any operation on the arrays causes...

Agreed here. If anything, I'd recommend eliminating the FIFO altogether and only accepting a single command (perhaps with a tag of the WEL bit status at the time the command...

I don't totally understand the use case, but maybe if we break this down a bit, a discussion will make it clearer. > my concern is that in the event...

Ah, you may have found one: That use case is ChromeOS's Closed-Case Debugging (CCD) setup, where the SBU1 and SBU2 can swap roles with the cable orientation. However, supporting pin...

I'm not sure what's going on with the Xcelium tests. They don't seem to actually run.

Note that DV pass rates are maintained. The last commit is there just to show that I've tested responses to multi-controller "fault" conditions while the target module was enabled. We...

Here's how the FPGA utilization numbers look compared to before the split FSM PR: | branch | flops | luts | | -------------- | ----- | ---- | | before-split...

I made a guess that Xcelium poorly handles dependency tracking between always_comb blocks and created an infinite loop that prevented forward progress with simulation. Let's see if moving `log_start` and...

> How did you check that the implementation works (at least in the common case)? Mostly, I made sure that DV regression test (+ top-level test) pass rates did not...