[clug] November CLUG Programmers' Special Interest Group meeting
darkstarsword at gmail.com
Mon Nov 8 20:58:13 MST 2010
> Abstract: Ian will possibly be talking about hacking the PowerPC
> to be little-endian. Confirmation of the talk and topic
> will be coming from Ian shortly after you read this
> half-baked email.
Ah, I knew I forgot to do something on the weekend, yes I'll be giving
Traditionally Linux on PowerPC has always been big endian, but some
PowerPC processors also support running in little endian mode, or
mapping certain pages as little endian. This is of particular interest
in the embedded space, where many graphics oriented SOCs and the
userspace code written to talk to them assume the whole world is
little-endian and therefore (without a lot of work) could not be
supported by the PowerPC architecture... until now.
Recently at OzLabs I created a proof of concept little-endian PowerPC
port of the Linux kernel, all the way to a busybox uClibc userspace,
and I'll be talking about my experiences doing so. There will be some
kernel code shown in the talk, though it will mostly be PowerPC
specific since the core kernel is already pretty endian-agnostic.
I'll go through the steps I followed to port the kernel, right from
the assembly to switch from big endian to little endian, through
getting the kernel to boot further and further, to the first squeal
from userspace as a hello world assembly program actually ran and the
final hurdles getting the rest of userspace to work (like the
compiler, toolchain, uClibc and buildroot).
There's also a short write up of my work on LWN:
See you all Thursday,
On the day *I* go to work for Microsoft, faint oinking sounds will be
heard from far overhead, the moon will not merely turn blue but
develop polkadots, and hell will freeze over so solid the brimstone
will go superconductive.
-- Eric S. Raymond, 2005
Please avoid sending me Word or PowerPoint attachments.
More information about the linux