On Thu, 30 Nov 2017 13:08:17 +1100
Alexey Kardashevskiy wrote:
> On 30/11/17 02:33, Alex Williamson wrote:
> > On Wed, 22 Nov 2017 16:16:49 +1100
> > Alexey Kardashevskiy wrote:
> >
> >> On some platforms INTx may not be enabled on a KVM host (one such
> >> example is IBM pHyp hypervisor and
On 30/11/17 02:33, Alex Williamson wrote:
> On Wed, 22 Nov 2017 16:16:49 +1100
> Alexey Kardashevskiy wrote:
>
>> On some platforms INTx may not be enabled on a KVM host (one such
>> example is IBM pHyp hypervisor and this is intentional). However
>> the PCI_INTERRUPT_PIN is not 0 so QEMU tries i
On Wed, 22 Nov 2017 16:16:49 +1100
Alexey Kardashevskiy wrote:
> On some platforms INTx may not be enabled on a KVM host (one such
> example is IBM pHyp hypervisor and this is intentional). However
> the PCI_INTERRUPT_PIN is not 0 so QEMU tries initializing INTx, fails as
> (!vdev->pdev->irq) in
On some platforms INTx may not be enabled on a KVM host (one such
example is IBM pHyp hypervisor and this is intentional). However
the PCI_INTERRUPT_PIN is not 0 so QEMU tries initializing INTx, fails as
(!vdev->pdev->irq) in the VFIO's vfio_intx_enable() and this is
a fatal error.
This adds a deb