If we need to inject an MSI into the guest, we rely at the moment on a working GSI MSI routing functionality. However we can get away without IRQ routing, if the host supports MSI injection via the KVM_SIGNAL_MSI ioctl. So we try the GSI routing first, but if that fails due to a missing IRQ routing functionality, we fall back to KVM_SIGNAL_MSI (if that is supported).
Signed-off-by: Andre Przywara <[email protected]> --- virtio/pci.c | 29 ++++++++++++++++++++++++++--- 1 file changed, 26 insertions(+), 3 deletions(-) diff --git a/virtio/pci.c b/virtio/pci.c index e9f36c7..04176c1 100644 --- a/virtio/pci.c +++ b/virtio/pci.c @@ -193,8 +193,20 @@ static bool virtio_pci__specific_io_out(struct kvm *kvm, struct virtio_device *v gsi = irq__add_msix_route(kvm, &vpci->msix_table[vec].msg); - if (gsi >= 0) - vpci->config_gsi = gsi; + /* + * We don't need IRQ routing if we can use + * MSI injection via the KVM_SIGNAL_MSI ioctl. + */ + if (gsi == -ENXIO && + vpci->features & VIRTIO_PCI_F_SIGNAL_MSI) + break; + + if (gsi < 0) { + die("failed to configure MSIs"); + break; + } + + vpci->config_gsi = gsi; break; case VIRTIO_MSI_QUEUE_VECTOR: vec = ioport__read16(data); @@ -205,8 +217,19 @@ static bool virtio_pci__specific_io_out(struct kvm *kvm, struct virtio_device *v gsi = irq__add_msix_route(kvm, &vpci->msix_table[vec].msg); - if (gsi < 0) + /* + * We don't need IRQ routing if we can use + * MSI injection via the KVM_SIGNAL_MSI ioctl. + */ + if (gsi == -ENXIO && + vpci->features & VIRTIO_PCI_F_SIGNAL_MSI) break; + + if (gsi < 0) { + die("failed to configure MSIs"); + break; + } + vpci->gsis[vpci->queue_selector] = gsi; if (vdev->ops->notify_vq_gsi) vdev->ops->notify_vq_gsi(kvm, vpci->dev, -- 2.9.0 _______________________________________________ kvmarm mailing list [email protected] https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
