Re: [PATCH 5.4 51/52] x86/hpet: Use another crystalball to evaluate HPET usability

[Jakub Kicinski]

On Mon, 11 Oct 2021 16:05:50 +0200 Greg Kroah-Hartman wrote:
> > FWIW I've never seen any problems prior to Paul's rework of bad clock
> > detection in 5.13. Backports to 5.4 and 5.10 are not necessary.  
> 
> Given that the hardware is still just as broken in those older kernels,
> why not?

Just filling in with extra context, I did say "FWIW" ;)

I don't use 5.4 nor do I understand the consequences of bad hpet 
well enough to comment on risk vs reward here. By consequences 
of bad hpet I mean whether its going to impact anything beyond
the tsc -> hpet fallback (which doesn't impact <5.13).