On 06/06/2019 19:08, Greg Kurz wrote:
> If KVM is too old to support XIVE native exploitation mode, we might end
> up using the emulated XIVE after CAS. This is sub-optimal if KVM in-kernel
> XICS is available, which is the case most of the time.
>
> Also, an old KVM may not allow to destroy and re-create the KVM XICS, which
> is precisely what "dual" does during machine reset. This causes QEMU to try
> to switch to emulated XICS and to crash because RTAS call de-registration
> isn't handled correctly. We could possibly fix that, but again we would
> still end up with an emulated XICS or XIVE.
Yes. I think we should favor a KVM device over an emulated one in any case.
>
> "dual" is definitely not a good choice with older KVMs. Internally force
> XICS when we detect this.
>
> Signed-off-by: Greg Kurz <gr...@kaod.org>
Reviewed-by: Cédric Le Goater <c...@kaod.org>
we could move the KVM tests done at the beginning of spapr_irq_init()
in routine spapr_irq_check() but this is for another patch, and it's
only a cleanup.
Thanks,
C.
> ---
> hw/ppc/spapr_irq.c | 10 ++++++++++
> 1 file changed, 10 insertions(+)
>
> diff --git a/hw/ppc/spapr_irq.c b/hw/ppc/spapr_irq.c
> index 3156daf09381..d788bd662a7a 100644
> --- a/hw/ppc/spapr_irq.c
> +++ b/hw/ppc/spapr_irq.c
> @@ -18,6 +18,7 @@
> #include "hw/ppc/xics_spapr.h"
> #include "cpu-models.h"
> #include "sysemu/kvm.h"
> +#include "kvm_ppc.h"
>
> #include "trace.h"
>
> @@ -668,6 +669,15 @@ static void spapr_irq_check(SpaprMachineState *spapr,
> Error **errp)
> return;
> }
> }
> +
> + /*
> + * KVM may be too old to support XIVE, in which case we'd rather try
> + * to use the in-kernel XICS instead of the emulated XIVE.
> + */
> + if (kvm_enabled() && !kvmppc_has_cap_xive() &&
> + spapr->irq == &spapr_irq_dual) {
> + spapr->irq = &spapr_irq_xics;
> + }
> }
>
> /*
>
