On Thu, 21 Nov 2019 11:56:02 +1100
David Gibson wrote:
> Due to the way feature negotiation works in PAPR (which is a
> paravirtualized platform), we can end up changing the global irq chip
> at runtime, including it's KVM accelerate model. That causes
> complications for VFIO devices with INTx,
On Thu, Nov 21, 2019 at 06:28:07PM -0700, Alex Williamson wrote:
> On Fri, 22 Nov 2019 12:18:24 +1100
> David Gibson wrote:
>
> > On Thu, Nov 21, 2019 at 09:57:38AM -0700, Alex Williamson wrote:
> > > On Thu, 21 Nov 2019 11:56:02 +1100
> > > David Gibson wrote:
> > >
> > > > Due to the way fe
On Fri, 22 Nov 2019 12:18:24 +1100
David Gibson wrote:
> On Thu, Nov 21, 2019 at 09:57:38AM -0700, Alex Williamson wrote:
> > On Thu, 21 Nov 2019 11:56:02 +1100
> > David Gibson wrote:
> >
> > > Due to the way feature negotiation works in PAPR (which is a
> > > paravirtualized platform), we c
On Thu, Nov 21, 2019 at 09:57:38AM -0700, Alex Williamson wrote:
> On Thu, 21 Nov 2019 11:56:02 +1100
> David Gibson wrote:
>
> > Due to the way feature negotiation works in PAPR (which is a
> > paravirtualized platform), we can end up changing the global irq chip
> > at runtime, including it's K
On Thu, 21 Nov 2019 11:56:02 +1100
David Gibson wrote:
> Due to the way feature negotiation works in PAPR (which is a
> paravirtualized platform), we can end up changing the global irq chip
> at runtime, including it's KVM accelerate model. That causes
> complications for VFIO devices with INTx,
Due to the way feature negotiation works in PAPR (which is a
paravirtualized platform), we can end up changing the global irq chip
at runtime, including it's KVM accelerate model. That causes
complications for VFIO devices with INTx, which wire themselves up
directly to the KVM irqchip for perform
