[SIG-ARM] Has somebody got interest in Raspberry Pi?

[Michael Tremer <michael.tremer@ipfire.org>]

[-- Attachment #1: Type: text/plain, Size: 2714 bytes --]

Hello,

it has been a little bit quite on this list, so I am going to start a
new conversation about Raspberry Pi, a cheap board with an ARM SoC.

Originally, there has been a request on the forums whether IPFire will
support this hardware. I would like to pass this question on to the SIG.

On the pro side of the equation:

The Raspberry Pi Foundation is a charity and is trying to bring small
PCs to poor countries and people who can not afford any. I'd like to
support that.

Usable software is still missing, but I guess the major distributions
will come up with such very soon. However, among them, there is a
router/firewall distribution missing, and as IPFire is the only Open
Source firewall which has support for ARM, we should consider to enhance
this support for the Raspberry Pis as well. We could bring firewalls to
poorer countries which is quite nice with IPFire: ISPs are planning to
provide LTE (which works with IPFire) and providing a connection to the
internet to small networks like in schools. I suppose that is what the
R-Pi-Foundation is aiming to do.

On the technical side: The hardware should be powerful enough for
connections to a couple of MBits/s but too weak for powerful services
like the filtering proxy and the intrusion detection system won't work
well.

The userland of the ARM port should run without any major modification.
Maybe we need to make some amendments on tools like the bootloader
(U-Boot) or others, but we should not expect any trouble there. The left
thing that is to do is to build a new kernel for that device. It uses a
Broadcom chipset which is supported by the Linux kernel and does not
need any patches unless we want to use the GPU.

The contra side:

The Raspberry Pi Foundation apparently messed up the launch. It is not
possible to get one and a lot of blokes are going to buy the few that
were already built to create home servers and stuff like that. That's
bad for us, because we cannot test any created code without the
hardware.

The hardware is weak. As mentioned earlier, it will work for a lot of
things, but not all of the features IPFire provides. I consider the
project rather a toy than a serious piece of computing hardware, so
supposedly the amount of interest will decrease soon when people realize
that they cannot really do what they intended to do with the boards in
the first place.

-----------------------------------------------------------------------

The question now is, if it should take the effort and build a new kernel
for the Broadcom SoC and support the Raspberry Pi boards. Is somebody
willing to do this? Has someone already tried something out? Did you try
to order a board?

Please mail me your thoughts.

Michael