> From: Jason Gunthorpe <jgg@xxxxxxxxxx> > Sent: Thursday, September 5, 2024 8:02 PM > > On Thu, Sep 05, 2024 at 08:29:16AM +0000, Tian, Kevin wrote: > > > Could you elaborate why the new uAPI is for making vPCI "bind capable" > > instead of doing the actual binding to KVM? > > I don't see why you'd do any of this in KVM, I mean you could, but you > also don't have to and KVM people don't really know about all the VFIO > parts anyhow. > that's not my point. I was asking why this VFIO uAPI is not for linking/ binding a vPCI device to KVM (not do it in KVM) while making it just 'bind capable'. 😊
