Known issue 20037708 (watchdog and redundant boots)

Hi, I’m looking through the release notes for l4t 28.2.1.

I found the known issue 20037708 where you recommend to disable watchdog.

I think I don’t fully understand the description, could you confirm if my understanding is correct?

  • in situation when preferred is slot with ‘_b’ postfix and system boots normally
  • if ‘’ postfix partitions get corrupted and ‘_b’-booted system encounters watchdog reset
  • during next start MB1 tries to boot from ‘’ partitions and it fails?

Do I understand correctly that MB1 has a feature to switch activities during resets caused by watchdog triggers? Can this be somehow disabled?

Since the issue seem to be in MB1, why did you release this for production? It seems it was known issue in the prerelease. Why can’t this be fixed in MB1 software (or is MB1 in silicon…?)?

Is there perhaps any external watchdog present on the TX2 board?

Is this being resolved in 31.0.1? If yes can you tell when is it going to leave beta?

Hi,

I’ve the same question. What is the status of this known issue - 20037708 from 28.2.1.
Is it safe to assume it is fixed in 31.0.2 since it is not listed as known issue anymore ?
Thanks

Issue 200377085 should be fixed in next rel-28 based release.

As for rel-31, we are still working on the fix.

Thank You WayneWWW. When is the next rel-28 release going tobe available and what would be the version I should look for - 28.2.2 or 28.3 or something else ? Thanks