On Tue, 8 Jun 2021 19:45:17 -0300
Jason Gunthorpe <jgg@xxxxxxxxxx> wrote:
On Tue, Jun 08, 2021 at 03:26:43PM -0600, Alex Williamson wrote:Hmm, that might have been a testing issue; combining driverctl with
Does it? How? The driver_override flag is per match entry not for thedrivers that specifically opt into this feature and the driver now hasIn doing so, this also breaks the new_id method for vfio-pci.
the opportunity to provide a proper match table that indicates what HW
it can properly support. vfio-pci continues to support everything.
entire device so new_id added things will work the same as before as
their new match entry's flags will be zero.
manual new_id testing might have left a driver_override in place.
By marking it a "vfio driver override"? :-\Sorry, with so many userspace regressions, crippling theOn the other hand it overcomes all the objections from the last go
driver_override interface with an assumption of such a narrow focus,
creating a vfio specific match flag, I don't see where this can go.
round: how userspace figures out which driver to use with
driver_override and integrating the universal driver into the scheme.
pci_stub could be delt with by marking it for driver_override like
But driverctl as a general tool working with any module is not reallyWe can't break userspace, which means new_id and driver_override need
Is the only issue the blocking of the arbitary binding? That is not a
critical peice of this, IIRC
to work as they do now. There are scads of driver binding scripts in
the wild, for vfio-pci and other drivers. We can't assume such a
narrow scope. Thanks,