Peter Memishian wrote:
> Netboot (legacy netboot, as opposed to DHCP) is on the way out. The
> little bit of code in genunix could easily be changed to use DLPI style
> 1, if the NICs all export DLPI style 1 nodes.
It's more complicated than that -- there are no nodes in /devices for the
DLPI style-1 and we're too early in boot to create them. Further, again,
no DLPI consumers and no GLDv3 drivers need to be concerned with DLPI
styles anyway because it's handled by libdlpi and the GLDv3 framework.
That little bit of code is only needed for "ce", and perhaps "ge". It
would be pretty trivial to make "ce" export DLPI style 1 nodes (and "ge"
as well -- heck "ge" could be trivially converted to GLDv3 since it uses
the same controller as eri, and that is already GLDv3.) All other
network bootable devices already (I believe) export DLPI style 1 nodes.
So, I don't think we need help from softmac here.
> > See previous discussion: if we update softmac to just be a shim (again,
> > this was the original design), then supporting these legacy drivers seems
> > very low-cost. Further, if we have to continue to support DLPI for
> > non-Ethernet (e.g., PPP) then what exactly does removing this simplify?
>
> Eventually removing the complexity of softmac itself! At least for
> Ethernet. (softmac doesn't do *anything* for PPP or other non-Ethernet
> links.)
Again, if softmac is properly factored it is merely something we remove
when we no longer need it. That's why we want to keep it as a separate
module and not weld it into mac. Yes, sins have been committed there --
mostly after the initial integration of Clearview UV -- and they need to
be undone.
Agreed... I want to clean it up. But part of the process needs to
include *notification* of the intent that these will be removed at some
point in the future.... so I'm asking about making the interfaces
Obsolete, not about actually removing them. Its *obviously* premature to
remove anything.
> I'm deep in this code now... and I *do* see value in cleaning this stuff
> up, at some point.
I'd rather see effort be put into areas that are straining under the load
and affecting customers, like GLDv3 buffer management.
I really want GLDv3 buffer management as well. But right now I'm hip
deep in this code specifically to fix a busted assumption (mac instance,
i.e. ppa == ddi_get_instance(dip)... this is a real problem for real
hardware we intend to be shipping.)
- Garrett
_______________________________________________
networking-discuss mailing list
[email protected]
