Re: [Xen-devel] xl: pci completion error

[Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>]

> > We could probably benefit from a boolean "force" parameter to do_pci_remove
> > and libxl_device_pci_remove, that would cause do_pci_remove not to
> > return if libxl__wait_for_device_model returns error.
> > Of course libxl_device_pci_remove would be called with force=1 by
> > libxl_device_pci_shutdown.
> > Gianni, what do you think?
> 
> Sergeys analysis sounds very plausible to me actually. The co-ordination
> required between qemu and libxl for PCI passthrough is very complicated
> and one ought to have fairly low confidence in it's correctness :P
> 
> Below is a less hacky version of the patch I just sent. Stefano, please
> consider this for inclusion.
> 

I like the patch, I am going to apply it.


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel