On Wed, Jan 27, 2021 at 02:28:13PM +0000, Dr. David Alan Gilbert wrote:
* Stefano Garzarella (sgarz...@redhat.com) wrote:
On Mon, Jan 18, 2021 at 04:03:12PM +0000, Dr. David Alan Gilbert wrote:
> * Stefano Garzarella (sgarz...@redhat.com) wrote:
> > Commit 9d7bd0826f introduced a new 'use-disabled-flag' property
> > set to true by default.
> > To allow the migration, we set this property to false in the hw_compat,
> > but in the wrong place (hw_compat_4_1).
> >
> > Since commit 9d7bd0826f was released with QEMU 5.0, we move
> > 'use-disabled-flag' property to hw_compat_4_2, so 4.2 machine types
> > will have the pre-patch behavior and the migration can work.
>
> Be a little careful that fixing this probably causes a migration from
> 5.2->6.0 to fail with this machine type; so when we do these
> type of fixes we often fix an old machine type between some pair of qemu
> versions and then break it between a different set.
Good point!
I did some tests using the example below always using pc-q35-4.2 and it
seems that works well:
- 5.2 -> 6.0 pass
- 5.2 -> 4.2 FAIL
- 6.0 -> 5.2 pass
- 6.0 -> 4.2 pass
- 4.2 -> 5.2 pass
- 4.2 -> 6.0 pass
Should I run some more tests?
Apologies for the delay; I had to step back and understand a bit about
what was going on.
no problem :-)
The problem here is that you're sending a 'disabled' subsection when
that option is true; your patch doesn't change the 4.1 machine type but
it does change the 4.2 machine type; and that makes the 4.2 machine type
not send it; so that means your patched version *will* work to existing
code (because it's a subsection anyway it doesn't break the stream
format when it's missing).
Thanks for the detailed explanation!
Reviewed-by: Dr. David Alan Gilbert <dgilb...@redhat.com>
Thanks,
Stefano
