lowRISC
[kmac] V1 Signoff
Description
Ensure V1 signoff criteria are fulfilled after focus area changes have landed.
While doing the PRNG work (#20820 ) I've run the full regression several times. Also the DV integration is done already for the new PRNG. I don't think that there is 12 days left of DV work.
All relevant DV work has been done together with the RTL work already (tracked in issues). I've thus lowered the effort estimate here to 0. FYI @johngt
kmac has seen the changes listed in #20978 since it was previously signed off at V2S. Those changes are covered as follows by DV:
- #19248 -- microarchitectural change covered by existing tests now that the assertion is no longer disabled
- #19262 -- microarchitectural change covered by existing tests
- #21624 -- architectural change covered by existing tests
- #21657 -- does not seem to be covered by the scoreboard --> created #22341 to track this for M4
- #21787 -- architectural change covered by existing tests
The currently open DV issues (also listed in #20978) are V3 items and one small V2S enhancement.
All V1, V2, and V2S tests pass with much more than 90% rate according to current nightly results, both for masked and for unmasked kmac.
All coverage results are well above 90% too:
Based on this analysis and results, I propose keeping kmac signed off at V2S.
@vogelpi WDYT?
Hi @andreaskurth , this looks good to me. There was only one change which isn't verified yet (the status type FIFO empty interrupt) but it also wasn't verified before (when it was of type event). Software currently doesn't use this interrupt anyway (because it wasn't really usable before).
LGTM!
Ah, and thanks for pulling the information together and preparing the issue @andreaskurth !