White/blacklisting manf_id/prod_id vs. func_id

[Dominik Brodowski]

On Tue, Nov 21, 2006 at 08:59:11AM +0000, Russell King wrote:
> On Tue, Nov 21, 2006 at 04:08:02AM +0100, Peter Stuge wrote:
> > On Mon, Nov 20, 2006 at 09:07:03PM +0000, Russell King wrote:
> > > However, it doesn't scale.
> > 
> > Feel free to suggest something that does, yet produces 0 false
> > positives.
> 
> Going back to what we had originally?  That seemed to work quite well
> and did not require a constant stream of ID additions to all the drivers.

Actually, it is equivalent what was there originally: if userspace is up and
running, devices with only one function and a valid func_id are matched. 
Others (including multi-function devices) require an ID -- but they already
did before; though to /etc/pcmcia/config instead of the driver.

> > I guess it comes down to deciding between white- or blacklisting.
> > Which of course just depends on what the common case for PCMCIA
> > devices is. Make sense?
> 
> The problem seems to be that even for _good_ devices, we're still
> having to add their IDs to the tables.  That's absolutely ludicrous.

That's not true.

> Good devices should just work out of the box as they used to with
> the original pcmcia-cs implementation.

They do work out of the box. They just might not work if userspace is not up
and running -- but that's not a regression. So please stop making this
claim.

> What are end users supposed to do when they buy new cards?  Build
> their own kernels?  Scream at their vendor to merge a patch which
> adds their card IDs to a driver?

What's the story with PCI drivers and devices?

One thing which might be worth thinking about is adding an interface similar
to PCI's new_id. Patches are welcome.

	Dominik