On Mon, 2025-05-05 at 10:54 +0200, Shalini Chellathurai Saroja wrote: > On 2025-04-28 14:05, Nina Schoetterl-Glausch wrote: > > On Thu, 2025-04-10 at 17:09 +0200, Shalini Chellathurai Saroja wrote: > > > Add Control-Program Identification (CPI) device to QOM only when the > > > virtual > > > machine supports CPI. CPI is supported from "s390-ccw-virtio-10.0" > > > machine > > > and higher. > > > > > > Signed-off-by: Shalini Chellathurai Saroja <shal...@linux.ibm.com> > > > --- > > > hw/s390x/s390-virtio-ccw.c | 10 +++++++++- > > > include/hw/s390x/s390-virtio-ccw.h | 1 + > > > 2 files changed, 10 insertions(+), 1 deletion(-) > > > > > > diff --git a/hw/s390x/s390-virtio-ccw.c b/hw/s390x/s390-virtio-ccw.c > > > index 7f28cbd1de..81832ee638 100644 > > > --- a/hw/s390x/s390-virtio-ccw.c > > > +++ b/hw/s390x/s390-virtio-ccw.c
[...] > > Fixing this in a separate commit could be bad for bisecting. > > Ok. > > > You introduce use_cpi in an earlier commit set to false and > > then flipping it in the migration patch for new machines. > > This way there is no broken intermediate state. > > > > I would also squash the compat migration changes into the previous > > patch. > > > > Hello Nina, > > If use_cpi is set to false in patch 1, then the sclpcpi device will not > be instantiated even for new machines at that point. The sclpcpi device > will only be instantiated when the use_cpi is set to true in the > migration patch. > > I prefer to squash this entire patch to patch 1, then the sclpcpi device > will only be instantiated for new machines with the code in patch 1 > itself and will not be dependent on the migration patch. I like this > approach as the logic to add sclpcpi device is complete in patch 1. > > What do you think?, thank you. Sounds good! [...] -- IBM Deutschland Research & Development GmbH Vorsitzender des Aufsichtsrats: Wolfgang Wendt Geschäftsführung: David Faller Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294
