Re: [PATCH v3 1/2] spapr: Disable legacy virtio devices for pseries-5.0 and later

[Greg Kurz <groug@kaod.org>]

On Tue, 10 Mar 2020 07:03:34 -0400
"Michael S. Tsirkin" <mst@redhat.com> wrote:

> On Tue, Mar 10, 2020 at 10:43:29AM +0100, Greg Kurz wrote:
> > On Thu, 5 Mar 2020 11:31:46 +0100
> > Greg Kurz <groug@kaod.org> wrote:
> > 
> > > On Thu,  5 Mar 2020 15:30:08 +1100
> > > David Gibson <david@gibson.dropbear.id.au> wrote:
> > > 
> > > > PAPR specifies a kind of odd, paravirtualized PCI bus, which looks to
> > > > the guess mostly like classic PCI, even if some of the individual
> > > > devices on the bus are PCI Express.  One consequence of that is that
> > > > virtio-pci devices still default to being in transitional mode, though
> > > > legacy mode is now disabled by default on current q35 x86 machine
> > > > types.
> > > > 
> > > > Legacy mode virtio devices aren't really necessary any more, and are
> > > > causing some problems for future changes.  Therefore, for the
> > > > pseries-5.0 machine type (and onwards), switch to modern-only
> > > > virtio-pci devices by default.
> > > > 
> > > > This does mean we no longer support guest kernels prior to 4.0, unless
> > > > they have modern virtio support backported (which some distro kernels
> > > > like that in RHEL7 do).
> > > > 
> > > > Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> > > > ---
> > > 
> > > Reviewed-by: Greg Kurz <groug@kaod.org>
> > > 
> > > FWIW, I could test the following:
> > > - allows a RHEL7 guest with pre 4.0 kernel to boot, as mentioned
> > >   in the changelog
> > > - breaks boot of older RHEL 6.10 guests as expected
> > > - allows migration of older machine types to/from QEMU 4.2
> > > 
> > > Tested-by: Greg Kurz <groug@kaod.org>
> > > 
> > 
> > Wait... I gave a try to virtiofsd and there's a problem:
> > 
> > $ ./ppc64-softmmu/qemu-system-ppc64 -device vhost-user-fs-pci
> > Unexpected error in object_property_find() at /home/greg/Work/qemu/qemu-ppc/qom/object.c:1231:
> > qemu-system-ppc64: -device vhost-user-fs-pci: can't apply global virtio-pci.disable-legacy=on: Property '.disable-legacy' not found
> > Aborted (core dumped)
> > 
> > It is still not possible to set the disable-legacy prop on the
> > vhost-user-fs-pci device, even without this patch, but QEMU
> > doesn't abort:
> > 
> > $ ./ppc64-softmmu/qemu-system-ppc64 -device vhost-user-fs-pci -global virtio-pci.disable-legacy=on
> > qemu-system-ppc64: -device vhost-user-fs-pci: can't apply global virtio-pci.disable-legacy=on: Property '.disable-legacy' not found
> > $
> > 
> > It seems to be related to the fact that vhost-user-fs-pci is a non-transitional
> > only device, as shown with this workaround:
> > 
> > --- a/hw/ppc/spapr.c
> > +++ b/hw/ppc/spapr.c
> > @@ -4574,7 +4574,7 @@ static void spapr_machine_latest_class_options(MachineClass *mc)
> >       * class since it doesn't have a compat_props.
> >       */
> >      static GlobalProperty compat[] = {
> > -        { TYPE_VIRTIO_PCI, "disable-legacy", "on", },
> > +        { TYPE_VIRTIO_PCI "-transitional", "disable-legacy", "on", },
> >      };
> >  
> >      mc->alias = "pseries";
> 
> Does this actually help? There's no type virtio-pci-transitional, is
> there?
> 

Oops you're right. There's no such type and it doesn't help in
the end... I should have tried that with -global before posting

   -global virtio-pci-transitional.disable-legacy=on

qemu-system-ppc64: warning: global virtio-pci-transitional.disable-legacy has invalid class name

Maybe worth adding a warning as well when done from the machine code, but
this is another story. Sorry for the noise.

> > 
> > but there's probably a better way to address this.
> > 
> > MST, Any suggestion ?
> 
> Hmm I'm not sure how to fix this properly. The only idea that comes
> to mind is a new internal-only "x-disable-legacy" that would be on virtio-pci,
> duplicating functionality of disable-legacy but intended for
> globals like this.
> 

Maybe but Daniel P. Berrangé jumped in and has some arguments
against what we're trying to achieve with this series...

A suivre.

> 
> 
> > > >  hw/ppc/spapr.c | 17 +++++++++++++++++
> > > >  1 file changed, 17 insertions(+)
> > > > 
> > > > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> > > > index 2eb0d8f70d..3cfc98ac61 100644
> > > > --- a/hw/ppc/spapr.c
> > > > +++ b/hw/ppc/spapr.c
> > > > @@ -65,6 +65,7 @@
> > > >  
> > > >  #include "hw/pci/pci.h"
> > > >  #include "hw/scsi/scsi.h"
> > > > +#include "hw/virtio/virtio-pci.h"
> > > >  #include "hw/virtio/virtio-scsi.h"
> > > >  #include "hw/virtio/vhost-scsi-common.h"
> > > >  
> > > > @@ -4566,8 +4567,20 @@ static const TypeInfo spapr_machine_info = {
> > > >  
> > > >  static void spapr_machine_latest_class_options(MachineClass *mc)
> > > >  {
> > > > +    /*
> > > > +     * Most defaults for the latest behaviour are inherited from the
> > > > +     * base class, but we need to override the (non ppc specific)
> > > > +     * default behaviour for virtio.  We can't do that from the base
> > > > +     * class since it doesn't have a compat_props.
> > > > +     */
> > > > +    static GlobalProperty compat[] = {
> > > > +        { TYPE_VIRTIO_PCI, "disable-legacy", "on", },
> > > > +    };
> > > > +
> > > >      mc->alias = "pseries";
> > > >      mc->is_default = true;
> > > > +
> > > > +    compat_props_add(mc->compat_props, compat, G_N_ELEMENTS(compat));
> > > >  }
> > > >  
> > > >  #define DEFINE_SPAPR_MACHINE(suffix, verstr, latest)                 \
> > > > @@ -4607,6 +4620,9 @@ DEFINE_SPAPR_MACHINE(5_0, "5.0", true);
> > > >  static void spapr_machine_4_2_class_options(MachineClass *mc)
> > > >  {
> > > >      SpaprMachineClass *smc = SPAPR_MACHINE_CLASS(mc);
> > > > +    static GlobalProperty compat[] = {
> > > > +        { TYPE_VIRTIO_PCI, "disable-legacy", "auto" },
> > > > +    };
> > > >  
> > > >      spapr_machine_5_0_class_options(mc);
> > > >      compat_props_add(mc->compat_props, hw_compat_4_2, hw_compat_4_2_len);
> > > > @@ -4614,6 +4630,7 @@ static void spapr_machine_4_2_class_options(MachineClass *mc)
> > > >      smc->default_caps.caps[SPAPR_CAP_FWNMI_MCE] = SPAPR_CAP_OFF;
> > > >      smc->rma_limit = 16 * GiB;
> > > >      mc->nvdimm_supported = false;
> > > > +    compat_props_add(mc->compat_props, compat, G_N_ELEMENTS(compat));
> > > >  }
> > > >  
> > > >  DEFINE_SPAPR_MACHINE(4_2, "4.2", false);
> > > 
>