Linux Kernel Network Subsystem Patching

[Lennart Sorensen]

On Tue, Feb 04, 2014 at 04:55:48PM -0500, D. Hugh Redelmeier wrote:
> | From: Lennart Sorensen <lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org>
> 
> | x32 is pretty neat.
> 
> It may be, but I have heard that there has been almost no take-up.
> I'd be interested to know if what I heard was wrong.

It hasn't been around very long, so I don't think that's relevant yet.

> My understanding is that 64-bit SPARC code is a lot less dense than
> 32-bit SPARC code.  x86-64 is almost as dense as x86 code.  Perhaps
> x32 is more dense that x86-64 or x86 code.

The pointer size still matters.

sparc 64bit still uses 32bit for each instruction, so other than the
pointer size doubling, the main issues is a few instructions can't be
used in 64bit code and take a few more instructions to replace (constant
address jumps I believe).  There doesn't seem to be any benefit to using
64bit on sparc in general though, unless you need the memory space.

-- 
Len Sorensen
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists