[Soc-coordination] GSOC project(s) for arm64 porting

Wookey wookey at wookware.org
Thu Apr 11 17:57:19 UTC 2013

+++ Steve McIntyre [2013-04-11 18:02 +0100]:
> Hi folks,
> I'd like to suggest a project (or maybe several) for the Summer of
> Code this year, working on arm64 (aka AArch64) porting. 

Good idea.

> As arm64 is a new architecture, there's quite a bit of work that could
> usefully be done by students in helping us to get the port going:
>  * setup of some infrastructure for building packages, both "natively"
>    in a software model and via cross-building

Within debian itself we have some core bits missing that need
(essentially) upstreaming from ubuntu if you want to actually build
anything (cross toolchain bootstrap, distro bootstrap, a lot of
multiarch patches). It should more or less 'just work' in experimental
with a bit of kicking but will still take someone a while.

Obviously individual package porting work is possible without actually
doing that (by building on either an openembedded or ubuntu raring
platform), but it won't be very 'debiany'.

>  * development of a test harness to compare the output of native- and
>    cross-builders, to help us verify that cross-building is working
>    correctly

This would be great.

>  * direct porting of software to the new architecture, targeting both
>    upstream development and Debian packaging (whichever is necessary)
> There are several DDs around who might be interested in mentoring
> projects in this area (me, Wookey, Riku, at least...) so we could run
> several projects for different students if that fits OK.

I'd be very happy to co-mentor for this (or even mentor if the focus
is on bootstrapping rather than porting). But I've already put in two
suggested projects (one bootstrapping tools, one cross-CI, the latter
of which would fit very well with your 'development of test-harness'
suggestion above, perhaps we should combine those two?). I don't want
to overcommit myself, but then how many slots do we have? I assume
it's not 20-something, which is the number of suggested projects we
have so far (including this), so not all of these are going to happen.x

Principal hats:  Linaro, Emdebian, Wookware, Balloonboard, ARM

More information about the Soc-coordination mailing list