[PATCH 1/2] PCI: Provide sensible irq vector alloc/free routines

[Christoph Hellwig]

On Fri, May 06, 2016 at 11:04:08AM -0500, Bjorn Helgaas wrote:
> I like this a lot.  One relatively minor comment below.

Do you want this addressed?  The problem is that I probably will also
need something like this flags argument for introducing the irq spreading
as we can't simply do it for all callers directly.  If you take
the patch as-is I'd be happy to do another pass later to remove it
and look into different no-MSI quirking once we're done with the
MSI-X spreading work.