On Sun, Nov 26, 2023 at 08:42:02PM +0100, Ferry Toth wrote:

> Funny discussion. As a hardware engineer (with no experience in automotive,
> but actual experience in industrial applications and debugging issues
> arising from bad shutdowns) let me add my 5ct at the end.

I suspect there's also a space here beyond systems that were designed
with these failure modes in mind where people run into issues once they
have the hardware and are trying to improve what they can after the fact.

> Now, we do need to keep in mind that storing J in a supercap, executing a
> CPU at GHz, storing GB data do not come free. So, after making sure things
> shutdown in time, it often pays off to shorten that deadline, and indeed
> make it faster.

Indeed.