[upstream] OpenTitan uart

[moidx]

Upstream OpenTitan uart model replacing ibex_uart.

[lexbaileylowrisc]

Hi, I've been assigned to look at this. CC @amit-kumarh

I have heard that you will eventually need more than just the uart, but that this is just the highest priority device you need? Is that correct?

I have rebased the latest opentitan branch on to upstream's master, and I'm trying to keep that up to date with changes. Currently this has introduced many errors. I am working on getting these errors resolved, and currently aim to get just enough working for the uart (since that's all this issue calls for) but this is requiring some effort to resolve and is ongoing. It currently builds fine, but crashes, so I'm looking for where I've introduced bugs in the rebase.

In the upstream qemu there is a machine called "opentitan", whereas in our fork there is "ot-earlgrey" and "ot-darjeeling".

My intent is that the "opentitan" machine will use the new uart. The new machine definitions (ot-earlgret and ot-darjeeling) already use the new uart, but are not yet upstreamed.

Does this suit your needs?
Is there any other specific details I should know about?

[amit-kumarh]

Hi Lex,

Thanks for the update!

The UART device is not necessarily the highest priority one for us - just one we (somewhat arbitrarily) chose as a starting point.

Starting with the pre-existing opentitan machine makes sense, in order to get the ball rolling on submitting patches.

We will eventually need a model that can simulate an EarlGrey device - is the plan to have the opentitan machine become this, or eventually upstream the ot-{earlgrey, darjeeling} models down the line?

[lexbaileylowrisc]

Okay, we will add the UART in to the opentitan machine. We will likely eventually transform the opentitan machine into an earlgrey implementation, but that will take much longer to complete.