[Hypervisor Live Update] Notes from April 21, 2025

[Mike Rapoport]

On Mon, Apr 28, 2025 at 05:35:07PM -0700, Dave Hansen wrote:
> On 4/27/25 23:20, David Rientjes wrote:
> > KHO v6 is now staged in Andrew's mm-new tree.  We discussed what it will
> > take for this series to be pushed to Linus, specifically around
> > Reviewed-by tags.  There is not a ton of x86 specific code, but it would
> > likely be useful to have Reviewed-bys from some x86 maintainers.  Dave
> > Hansen, would you be the right person to take a look at this from an x86
> > perspective?
> 
> It looks mostly OK.
> 
> The one worry is the scratch reservation sizing, how folks will get that
> right, and the implications if they get it wrong.

For engineered cloud systems folks will get it right most of the time with
explicit setting of the scrach size in the command line.
There's also a way to let kernel estimate the scratch size by scaling early
memory usage by user defined percentage (with 150% as the default).

> If it's undersized because newer kernels need more space, what ends up
> happening? Is it just an OOM during boot?

It will OOM and panic during boot, yes.
Potentially we can skip KHO if the scratch is undersized and continue to
boot without restoring any of the previous state, but to start with OOM and
panic are much simpler.

-- 
Sincerely yours,
Mike.