[test-triage] rom_e2e_smoke has started failing consistently about opentitan HOT 5 OPEN

Hi @rswarbrick, I can take this one for now since I have to clean-up all the rom e2e tests after switching to ECDSA. #21204

from opentitan.

Hmm. I've just tried running this with a known-good SHA (4ee21f8, which was used for the nightly on 28th March. The resulting log file has the same warning, so maybe that wasn't a smoking gun after all.

from opentitan.

Looking more closely at a failing log, the failures appear after this message:

Job chip_earlgrey_asic-sim-vcs_run_default killed due to: Exit reason: User job exceeded runlimit: User job timed out has 14 failures:

I suspect this means that we're just running out of time in the tests and getting killed by the runner. This might have nothing to do with this particular test at all. :-(

from opentitan.

Hmm, I can't remember why I made that change to continue if the log Db file was not present, perhaps it was because we were not generating one for each SW image at the time? but maybe we are now... that part probably requires more investigation but like you mention I don't think that is the reason `rom_e2e_smoke is failing.

I think the reason this failing is possibly due to a sigverify issue, I think it makes sense to just leave this open for a bit and fix it in a couple weeks. @moidx is currently breaking things as he:

1. converts sigverify to use ECDSA (instead of RSA), and
2. moves first stage secure boot keys from ROM to OTP.

from opentitan.

I think he is tracking all of this in #22697.

from opentitan.