Re: [PATCH v6] hw/i386: pass RNG seed via setup_data entry

["Jason A. Donenfeld" <Jason@zx2c4.com>]

Hi Paolo,

On Thu, Jul 21, 2022 at 01:47:02PM +0200, Paolo Bonzini wrote:
> On 7/21/22 13:00, Michael S. Tsirkin wrote:
> > Well why not.
> > 
> > Reviewed-by: Michael S. Tsirkin<mst@redhat.com>
> > 
> > who's merging this? Paolo me or you?
> 
> I don't think this should be merged as is.
> 
> The linuxboot ROM takes the data from fw_cfg, and (with the exception of 
> ACPI tables) that data is not migrated.  Because reading it into the 
> guest is not atomic, both sides must match.  This version of the patches 
> at least doesn't move the preexisting DTB entry of the setup_data, but 
> it still has a mismatching size and that can be a problem when migrating 
> backwards.

As discussed online, this seems absolutely preposterous and will never
happen anywhere real ever at all. Trying to account for it is adding
needless complexity for no real world benefit; it's the type of thinking
that results in a mess. Further, conditionalizing the RNG seed on
something else means fewer users receive the increased security of
having an early boottime seed. This seems like a silly direction go go
in.

But to assess things in the open here:
- On upgrades, there's no problem because the old bytes don't move.
- On downgrades, there's mostly no problem because next will point to 0.
- On downgrade there could be some ridiculous theoretical problem if the
  reader has already read a non-zero next. But this will never actually
  happen in practice.

So we really should just stick with the simple and straight forward path
that this v6 accomplishes, and not muck things up with stupidity.

Jason