Hi, I'm currently trying to make vhost-net work on ARM cortexA15 Arndale board. As irqfd is not yet upstream, I've used irqfd and irq routing patches recently posted on kvmarm list. So far I can properly generate irqs to the guest using vgic and irq-routing , but as far as I can see virtio registers (interrupt status register @ offset 0x60) are not updated by vhost kernel backend. Does it mean that virtio virtual registers have to be managed by Qemu, even when using a vhost kernel back-end ??? I cannot find inside vhost code where those registers could be updated, and I can clearly see that mmio accesses from guest to those registers causing a qemu virtio_mmio_read or virtio_mmio-write.
I thought vhost was added to prevent any qemu interaction in the data path . Could someone more familiar with vhost could tell me if and where I'm wrong ? Thanks a lot in advance Best regards Rémy -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
