Re: kern.boottime drift after boot?

[RVP <rvp%SDF.ORG@localhost>]

On Thu, 12 Oct 2023, Simon Burge wrote:

> Assuming of course that the boot_id is generated after interrupts
> are enabled and some jitter has been processed :).

That's almost certainly the case: that UUID is only generated when
userspace first reads it (and there aren't any kernel-consumers of
that UUID that I'm aware of). See: proc_do_uuid():

https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/drivers/char/random.c?h=v6.5.7#n1609

The file was put in to support server-farms where you push an update
to a bunch of machines, command them to reboot, then check if the
machines have actually rebooted for the updates to "take". But,
as abs@ just scoped out, it can have other uses too...

-RVP