Garrett D'Amore wrote: > DLPI style 2 is standards conformant, but DLPI style 1 is also part of > the same standard. Standards conformant programs are required to test > both DLPI style 1 *and* style 2. This is true with vanity naming anyway > -- Vanity names don't work with programs that *only* support style 2.
I don't actually see such a requirement in the standard, but since it's not really the point, I'll just drop that issue. The real point is that there are applications out there that assume the use of only Style 2. Heck, Sun's own snoop supported only Style 2 until fairly recently. It was very wide-spread exactly because Sun didn't bother shipping any Style 1 devices. I suspect the same was true of several other SysV vendors. And because Sun didn't bother shipping such devices, there was no way for any real application writer to code for Style 1 -- he can't reasonably test it, and what isn't tested obviously does not work. As I mentioned before, those Style 1 nodes are quite new compared with the bulk of software out there. Unless the assumption is that either (a) binary compatibility is "optional," (b) we just don't care if hidden flaws like that are exposed, (c) everybody can just recode and recompile old applications when they break, or (d) somehow the gains are much greater than the risk, I don't see how this is a viable plan. > It will make sure that developers are *not* relying *solely* on style 2 > to access the network devices. (Indeed, they should be using libdlpi, > but even if they don't, they need to at least start by trying a style 1 > open first.) I'd support any reasonable means to let developers know that supporting only Style 2 is a dumb idea, even if it appears to work, and even if there aren't any Style 1 devices on the (pre-S10) machines you're using. But I'm much less excited about obsolescence. > The proposal wasn't to kill all use of Style 2. The proposal was to > kill the guarantee that *all* drivers provide (via GLDv3) style 2. If a > pure DLPI driver wants to supply style 2, fine. I don't care what driver writers do. I care that applications that worked yesterday still work tomorrow. > But consumers *need* to > be smart enough to check for style 1. Then we can stop going through > the gyrations required to support style 2 in our drivers when really all > they ought to need to do is supply style 1. They "need to be" that smart, but at least until S10, there really was no way for them to be that smart. -- James Carlson 42.703N 71.076W <[email protected]> _______________________________________________ networking-discuss mailing list [email protected]
