On Wed, Apr 14, 1999 at 10:26:54AM -0700, Jonathan Stone wrote:
> 
> The comment says it puts the card into the 'wait for key' state, which
> is just what the code in isapnp_init() says its doing.
> And isapnp_isa_attach_hook() calls isapnp_init()...
> so it should be safe to put the cards into `wait for key' state
> the same way that Linux does.
> 
> I'm guessing that the isapnp_isa_attach_hook is also re-enabling
> response to the 3c509-specific ISA auto-scan.  I just dont know how
> this might mess up a later isapnp probe, if it turns out hte board was
> configured to pnp mode after all.
> 
> to put it another way: we need to test kernels configured for both isa
> and isapno in two cases:
> 
>    1) card set up for isapnp by vendor firmware/bios
>    2) card not set up for isapnp

     3) other isapnp cards for which the isapnp_isa_attach_hook() was needed
	(I think it was the soundblaster, and another one too).

> 
> What I suggested should (may?) fix #2, but may also break #1.
> 
> (Manuel: can you try a patch from where you are today?)

So, sorry. I don't have physical access to the machine, and starting from
tomorow I'll be 500Kms away from it.

--
Manuel Bouyer <bouyer@antioche.eu.org>
--