On Tue, 2010-02-02 at 21:08 -0500, Stefan Monnier wrote: > While I'm not convinced that there is still a need for a patch like > drivers-ata-ata_piix-postpone-pata.patch (whose name seems misleading > BTW), seeing how most/all other distributions don't bother with such > a thing, I'm willing to believe that there might still be occasional > situations where the two drivers could still get into a race. Having two drivers for the same PCI device id does not result in a race condition. udev + modprobe will load them both, in unspecified order. Either the first will bind to the device, or if its probe function returns an error then the second will bind. > But I'd argue that nowadays, the patch should then be replaced by > another one. More specifically, I think that for those chipsets where > both ata_piix and piix can be used, we should nowadays prefer ata_piix > over piix, because it's better maintained. And for us docking users, it > offers very useful additional features. This is what we are going to do for squeeze. Ben. -- Ben Hutchings I'm not a reverse psychological virus. Please don't copy me into your sig.
Attachment:
signature.asc
Description: This is a digitally signed message part