[fedora-arm] activity

Manas Saksena msaksena at marvell.com
Fri Nov 16 21:22:36 UTC 2007

> From: fedora-arm-bounces at redhat.com
[mailto:fedora-arm-bounces at redhat.com] On Behalf Of Paul Pescitelli
> Sent: Friday, November 16, 2007 12:52 PM
> To: fedora-arm at redhat.com
> Subject: [fedora-arm] activity

> there does not seem to be much activity here...is this still alive ?

Hi Paul, 

Thanks for asking. There is plenty of activity, just not so much
in the foreground. I am hoping that in the next few weeks this 
will change significantly. I have some details below, and hopefully
this gives you a sense where things are headed. 

Much of the effort is in migrating to F8 and building on top of 
it. Hopefully, we will have a fully usable F8 based repository 
before the end of the year. 

The Fedora secondary architecture stuff is also coming together
and with luck we will be part of the F9 release as an official
secondary architecture. 

Fedora-8 ARM Repository

We have been building Fedora-8 packages for a while, and they
are available. There are two major issues facing us right now. 

1. The debuginfo packages are busted because of an issue with 
2. Glibc 2.7 does not build for ARM today. The packages today
   are built with glibc 2.6. 

Derivative Distributions

A key goal of this project is to create derivative distributions
based on the Fedora repository. Much of our effort is focused
on that front. Some of the efforts on that front. 

Fedora GIT Repository

We have an effort to convert the Fedora CVS into a git repo.
With a git repository, it is easier to manage modifications
to packages that are needed for "downstream" derivative
distributions. You can see Lennert's post on this topic to 
fedora-devel a few days back. 


Custom Root File Systems

One of our goals is to support the ability to create custom
root file systems to match your needs. 

The tool we are currently using to build these is "pilgrim". 
Once the F8 repository is available, we should have a few of
these available, and these can then become the starting point
for others. This is the same tool that is being used by OLPC
for creating the XO images. 

Small Footprint Root File Systems

While on desktop/server end, people don't worry about bloat
as much, on much of the embedded end, you do. So, part of the
effort is to make it easier to build custom root file systems
that are not bloated. 

One of the idea there is to not use SysVinit, coreutils, bash
etc. (which pull in a lot of dependencies) and replace them
with busybox. You can then easily build root file systems for
wireless access points and other types of devices and not 
gobs of space. 

The other effort there, to squeeze things down to even smaller
footprints is to use uclibc as the C-library. We have an effort
on that front as well. 

Cross-Compiler Toolchains

While we are doing native builds today, cross compilers and
cross-building of packages is useful for a number of reasons. 
So, we have been spending time on that front too. This allows
(for e.g.) to build a uclibc based system, or to bootstrap 
for a new ARM architecture variant (e.g., ARMv6) and so on. 

Best Regards, 

More information about the arm mailing list