On Thu, 3 May 2001, Bill Studenmund wrote:

> > As I said, I'm afraid if we start to provide such a pci_scan function,
> > it will be used, and that would be catastrophic from the architectural
> > cleanlyness POV imho.
> 
> Is architectural purity worth pciide drivers not working?
> 
> From my understanding of Manuel's comments, there are pciide chips out
> there where the vendor didn't bump the version of the pciide component
> when changing it. The only way to correctly deal with these chips is to
> look at the version of another pci device, and act accordingly.
> 
> So what do we do?

Oh, one thing I think I should mention is that I think all of the "pure"
elements Matthias is describing sound good, and we should do them. My main
point I wanted to make above is that I don't think that we should block
the addition of the scan routine just because we don't have all the pci
framework we would want for pci lkms.

Take care,

Bill