* [EMAIL PROTECTED] [2006-04-05 10:06:34]
> If I wanted to write a new virtual NIC for Solaris, what approach
> would I take?

GLDv2 is the recommended approach.  If for some reason GLDv2 doesn't
fit with what you want to do, you should write a DLPI driver.  If
neither GLDv2 nor DLPI fit with what you want to do then Solaris
probably doesn't have supported stable interfaces to help you.

> If I want to make a new NIC to experiment with a protocol like CARP
> or VRRP, where would I go for how to start on such a project?  Do we
> have sufficient interfaces to support this?

Is there something specific about CARP or VRRP that would require
special features of a NIC driver?

> Do I have to write a complete NIC driver, like a bge, even though
> gld is available, or can we make something else, better?

When originally created, bge was a GLDv2 driver, hence I'm confused by
"like a bge, even though gld is available".

Admittedly, bge is now a Nemo driver and I agree that the Nemo
interfaces should become stable, but Sun takes the stability of
interfaces very seriously, which makes anyone with some doubt about
the long-term validity of an interface very leery of declaring it
Stable.  Nemo is still undergoing rapid and significant change - today
may just not be the right time to put a line in the sand.

> Ideally I'd expect the IPMP code to be a model of how to do this.

IPMP is, in reality, a strange beast.  Why you'd expect it to set any
kind of precedent for how to write NIC drivers is something of a
mystery to me, as IPMP operates almost entirely at layer 3.  Hiding
some of the implementation details of IPMP, as the Clearview project
will do, seems like a good thing and isn't really related to the
stability level of NIC driver interfaces.

dme.
-- 
David Edmondson, Solaris Engineering, Sun Microsystems.

_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to