Date: Sun, 21 Sep 1997 13:34:19 +0200 (MET DST) From: Luigi Rizzo <luigi@labinfo.iet.unipi.it> To: mark@grondar.za (Mark Murray) Cc: jonny@coppe.ufrj.br, multimedia@FreeBSD.ORG Subject: Re: Report back on Luigi's sound driver and GUS PnP Message-ID: <199709211134.NAA14659@labinfo.iet.unipi.it> In-Reply-To: <199709210743.JAA14985@greenpeace.grondar.za> from "Mark Murray" at Sep 21, 97 09:43:27 am
next in thread | previous in thread | raw e-mail | index | archive | help
> I think what folk meant above (I certainly did :-) ) was that if a > device is declared such: > > device pcm0 > > then it should (?) itself from the PnP settings, not creating a "dead" > pcm0, and the "live" one from pnp1. > > Are you saying that if you have a > > device foo0 > > Where foo's were previously available in the ISA bus, then the first > workable fooN is foo1? YUKK! the line "device pcm0 at isa? ..." means that I _might_ have a legacy isa device as pcm0, and since PnP is probed and attached _before_ legacy isa devices, I cannot use the 0 entry for pnp devices. What I could do is to attach the pnp device, number, and when the isa probe fails shift the pnp unit to become unit 0. But this is quite messy, in the boot log you would have something like: pcm1 attached.... pcm0 not found. pcm1 has now become pcm0! if the config line as just "device pcm0" (without "at isa? ...") as you propose, perhaps things could be worked out (but then you could _not_ have legacy ISA device with that kernel). "device de0" is such an example since the de driver does not work with ISA devices, only PCI. So I consider the current setup the lesser of the two evils. if you find a better way please go ahead and send me the patches. Cheers Luigi
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199709211134.NAA14659>