Re: [RFC PATCH 12/12] ixgbevf: Kconfig, Makefile and Documentation

[Simon Horman <horms@verge.net.au>]

On Tue, Dec 08, 2009 at 02:49:21PM -0800, Rose, Gregory V wrote:
> >-----Original Message-----
> >From: Randy Dunlap [mailto:randy.dunlap@oracle.com]
> >Sent: Tuesday, December 08, 2009 2:42 PM
> >To: Kirsher, Jeffrey T
> >Cc: netdev@vger.kernel.org; gospo@redhat.com; Rose, Gregory V
> >Subject: Re: [RFC PATCH 12/12] ixgbevf: Kconfig, Makefile and
> >Documentation
> >
> >> +The ixgbevf driver supports 82599-based virtual function devices that
> >can only
> >> +be activated on kernels with CONFIG_PCI_IOV enabled.
> >> +
> >> +The guest OS loading the ixgbevf driver must support MSI-X
> >interrupts.
> >
> >Hi,
> >
> >Could the Kconfig language express some of these requirements, or
> >am I confused between the host and the guest OS?
> [Rose, Gregory V] 
> 
> The guest OS must also support MSI-X interrupts.  It is a requirement of our VF device to use MSI-X.
> 
> >
> >
> >> diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
> >> index 0bbd5ae..a484415 100644
> >> --- a/drivers/net/Kconfig
> >> +++ b/drivers/net/Kconfig
> >> @@ -2614,6 +2614,27 @@ config IXGBE_DCB
> >>
> >>  	  If unsure, say N.
> >>
> >> +config IXGBEVF
> >> +       tristate "Intel(R) 82599 Virtual Function Ethernet support"
> >> +       depends on PCI
> >
> >	depends on PCI_IOV && PCI_MSI
> >??
> >
> [Rose, Gregory V] 
> 
> The guest does not need to have IOV support configured to use the VF device but it does need MSI-X.  We'll incorporate this suggested change.
> 
> Thanks for your comments.

The igbvf driver can also run on the host or indeed on a system
without virtualisation. Is the same true of the ixgbevf driver?