Wei Liu

> Update: > > I've done away with the magic IO port and weird init sequence and restructured the device and driver to be PCI devices. I've never worked with...

> > > Update: > > > I've done away with the magic IO port and weird init sequence and restructured the device and driver to be PCI devices. I've...

> Regarding the allowing concurrent accesses change, other devices would have to handle their own locking to see improvements. Are you referring to virtio-net multiqueue? Yes, virtio-net.

> > > Regarding the allowing concurrent accesses change, other devices would have to handle their own locking to see improvements. Are you referring to virtio-net multiqueue? > > >...

@Dummyc0m I have a hack https://github.com/liuw/cloud-hypervisor/commit/37eba08b8036019dda3a7ab3e3aa3f5cc49d62da if you want to see the difference. This change forces virtio devices to go through the dispatching logic inside Cloud Hypervisor.

@Dummyc0m the code to CH is merged. Can we close this issue now?

A few comments: 1. I think this should be gated by a flag and be disabled by default, because kernel code is not yet upstreamed. 2. I think you should...

@novakovic please don't push to the existing branch like that. The top commit you pushed is not signed off. It looks like you're making a minor change in numbering. You...

I have a small patch to add a new build test. I can post that once this is merged.

Clippy (beta) is complaining, but those issues aren't introduced by this series.