On Tue, 2007-08-14 at 11:49 -0700, Artem Kachitchkine wrote:
> > Why is it ugly to extend the binding mechanism to allow greater
> > differentiation between devices?
>
> Binding mechanism is a direct reflectin of the bus's device
> identification mechanism. For the PCI bus it's the configuration space.
> PCI nexus can read config space registers and create compatible names
> without leaf drivers' involvement. Garret described devices that do not
> differentiate themselves via standard config space, but via other
> non-standard registers, which only leaf drivers know about. The driver
> binding process requiring driver assistance is sort of like Baron
> Munchhausen lifting himself from a swamp by pulling on his own hair.
Yes, but it does happen.
Other FOSS' do this by allowing their probe(9e) equivalent routine to
"fail" and another driver to try the same node.
I think being able to change the device node name, via "probe" or some
other mechanism, would be more elegant.
-- Garrett
>
> -Artem
> _______________________________________________
> opensolaris-code mailing list
> [email protected]
> http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
