Re: [PATCH 09/10] Change table chaining layout

[David Miller <davem@davemloft.net>]

From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Mon, 22 Oct 2007 14:47:38 -0700 (PDT)

> On Mon, 22 Oct 2007, Alan Cox wrote:
> 
> > Still doesn't answer the rather more important question - why not just
> > stick a NULL on the end instead of all the nutty hacks ?
> 
> You still do need one bit for the discontiguous case, so it's not like you 
> can avoid the hacks anyway (unless you just blow up the structure 
> entirely) and make it a separate member). So once you have that 
> bit+pointer, using a separate NULL entry isn't exactly prettier. 
> 
> Especially as we actally want to see the difference between 
> "end-of-allocation" and "not yet filled in", so you shouldn't use NULL 
> anyway, you should probably use something like "all-ones".

Indeed that's the crux of the matter, we need to express a trinary
state "end of scatterlist, next entry is linear, next entry is
indirect" plus a pointer for the indirect case.

Generally, Jens was doing a good job cooking up the patch that
implemented this fully and I took care of making sure tricky
ports like sparc64 built cleanly etc.

He went away for a few days, but when he gets back we should seriously
work on integrating his work.

I fully recognize Alan's m68k on-stack alignment concern.  The on-stack
cases are troublesome in other ways as well, and I think therefore the
way to move forward is to convert those to some kind of dynamic scheme.
Usually such code is working in a locked context on some object
(crypto instance, for example) and thus the scatterlist chunk can
be embedded into that object for ensured alignment.