Questa Sim 2020.01 Regressions about axi HOT 7 CLOSED

Thanks. I am reluctant to add a version-specific ifdef into our code because it increases complexity with very narrow benefit. So please fix this locally for now.

Let's collect regressions in Questa 2020.1 in this thread. I expect we find a pattern. That would allow us to file a bug report to Mentor or implement an efficient workaround in our code.

from axi.

When you comment those lines out, it simulates fine?

from axi.

Affirm

from axi.

I am observing a major bug in an axi_mux instance in Questa 2020.1.
All outgoing master request handshake signals are X - but I can observe their direct source (spill registers) as being non-X.
Interestingly, it happens only in one of two instances in the design; it might have something to do with the spill registers which only that instance has, I will investigate further today.
Edit: disabling (and also manually coding out) spill registers made no difference.

from axi.

I am observing a major bug in an axi_mux instance in Questa 2020.1.
All outgoing master request handshake signals are X -

Thanks for reporting, this sounds very serious!

but I can observe their direct source (spill registers) as being non-X.
Interestingly, it happens only in one of two instances in the design; it might have something to do with the spill registers which only that instance has, I will investigate further today.
Edit: disabling (and also manually coding out) spill registers made no difference.

Did you find out more about the failure mechanism?

from axi.

Did you find out more about the failure mechanism?

Not yet - I'm currently working on reproducing it in a more isolated setup. I'll report back if I find anything!

from axi.

At the integration meeting today we were running most of the AXI infrastructure on QuestaSim 2020.1 with no problems. I am starting to wonder whether there might be a problem with our internal installation.

from axi.