What is the difference between l15_header_ack abd l15_ack

[Yuda Wang]

Hi, 

I can tell that header_ack is sometimes ahead of ack, is there a mechanism? l15 is a private buffer, correct?

Best,

redpanda

[Jonathan Balkind]

Yeah this is a weird detail inherited from the OST1 core that we kept to keep things compatible but will probably kill later.

My recollection is you hold valid until header_ack, at which point you don't need to hold the values for your other request wires, but you need to keep valid held until _ack? Or maybe it's the other way round... It should be possible to infer from the Ariane or BlackParrot (in open PR) L1.5 wrapper code. There's basically only one or two cases where the precise timing causes an issue if you just hold val and your wires until ack, and one of those that I recall is for using the store interface to send interrupts on the NoC.

Thanks,

Jon

--
You received this message because you are subscribed to the Google Groups "OpenPiton Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpiton+...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/openpiton/4119375d-772a-44ce-999b-e94e5a697371n%40googlegroups.com.

[Yuda Wang]

Thanks for the clarification. Another question is how to see the 9000 test cases? I finished summarizing the monitors(l15, l2, jtag, manycore_networks...) recently.

Best,

Redpanda

To view this discussion, visit https://groups.google.com/d/msgid/openpiton/CAJm4tWZ42u88rbo4OLRMKU9dL_uM%2BqfYktA_GTe%2B8Sn74H7MjQ%40mail.gmail.com.

[Jonathan Balkind]

It might be easier to join our zulip chat (openpiton.zulipchat.com) to ask some of these questions.

The test cases are all under piton/verif/diag in various directories. I think the 9000 count is about the runnable regressions rather than the number of tests. If I recall the number of test files was around 3000-4000? It's been a long time so I could be off on that. Also it's less relevant today as the OST1 core isn't a priority any longer.

Thanks,

Jon

[Yuda Wang]

OK, Zulipchat registered. I will go over the test cases next week.

Best,

Redpanda