On 5 Dec 2007, at 14:49, Sebastien Roy wrote: > I noticed that you've become the RE on 6637329 which was filed > yesterday. I also noticed that the verbiage in the CR description > mentions that something will go looking for DDI_NT_NET nodes in the > devfs device tree: > >> dynamic updates to HAL The HAL network module at boot time >> attaches the Solaris device >> tree and creates nodes for each device exporting DDI_NT_NET >> device minor nodes. The module (via hald) then registers with >> sysevent a handler which dynamically updates the HAL tree in >> response to updates to the Solaris device tree. > > If the intent is to have HAL represent datalink interfaces and not > "networking devices", then the above strategy might be missing the > mark given the upcoming /dev/net filesystem and datalink interface > vanity naming. I'm not sure if you've already thought about how > this will work, but from my cursory look at the CR, it seems like > some interaction with the implementation of /dev/net might be > necessary.
I think that you're right. There's a pre-vanity naming prototype (by Artem) that resulted in the text you see in the bug. Thinking about the interaction with vanity naming is something that I've yet to do. dme.
