On Wed, Mar 14, 2018 at 01:35:19AM -0700, Christoph Hellwig wrote:
> Should this logic go into a little helper so that everyone is kept
> in sync?
Great point, thanks! What I'd really like is to completely get rid of
most of the pcibios_enable_device() implementations. Most of them
contain nothing that's arch-specific.
I'll look at that some more, but will probably do that as additional
steps on top of this one. That way this patch stays mostly trivial
and obvious as a generalization of 16cf0ebc35dd ("x86/PCI: Do not use
interrupt links for devices using MSI-X") to other affected arches.
Bjorn
