Fedora ARM F-17 and rawhide status update
or as the saying goes… the mustard indicates progress! Well there certainly has been an a-bun-dance of movement in the Fedora ARM world of late and it’s been way too long since I’ve given an update on the status of the Fedora ARM SIG.
We’re progressing well and there’s a lot of stuff going on. Of course you’re all going to be up be up at 06:00 GMT tomorrow (Wed) to order your Raspberry Pi running Fedora 14… aren’t you! I’m sure those of you that got a voucher for one at FUDCon Blacksburg will be ready and waiting.
In terms of raw package building of rawhide and F-17 branches we’ve been progressing well since we kicked off the mass rebuild shortly after the aforementioned FUDCon. In terms of numbers of a total 11348 packages in F-17 we’ve built 8861 of them giving us around 2487 left to go. Of course that’s not an exact number as there’s a number of x86 only packages and between now and the final beefy F-17 release there will be lots more releases of the packages that will need to be built. We’re already building rawhide to ensure we stay up to speed on that front too! One thing that is slowing us down is having to spend a lot of time fixing broken upstream packages and of course fixing ARM specific issues.
In terms of devices that will be supported for the Fedora 17 release the number will be quite small. The list at the moment will likely be PandaBoard (including ES), all modern BeagleBoards, Trimslice, Efika SmartTop and SmartBook, Raspberry Pi and the OLPC XO 1.75. The last two will be Fedora re-mixes. In most cases of pure Fedora the support will likely be console based or accelerated 2D UX. The device market and associated support is moving pretty quickly at the moment so that could easily change between now and release, this is very much a moving target. It’s also not to say that Fedora 17 won’t run on all devices, it’s just that Your Mileage May Vary and it depends a lot on the state of the mainline kernel support.
So what else are we working on? We’ve been doing a lot of planning in preparation for the Proposal to primary arch. We’ve also been working with various tools teams like Anaconda to get some features for building images. In this regard we also need to work with the appliance-tools and livecd-tools teams to see what can be done to allow them to support image creation. It would be cool to be able to use the appliance tools to create an qemu image that can be directly reported into libvirtd and virt-manager (hint hint ). I’m also slowly starting to investigate what’s needed to support DeviceTree so we can build less kernels and support more devices but I’m not sure it’s all quite there yet.
So what can you do to help? The most simple and best thing is to ensure your package has been built properly for mainline as part of the mass rebuild, or if we file a RHBZ bug with ARM issues to fix it or assist us in fixing it quickly. Most people have been fabulous with quick fixes and attention on bugs. Lastly we should have some minimal rootfs files available soon for F-17
In summary we’ve got nearly nine thousand individual packages already built for ARM supporting both ARMv7 and ARMv5tel architectures. For our Fedora 17 release will be on a reasonable number of innovative and interesting devices. We’re aiming to release as close as is humanly possible to the Mainline Fedora release. Users should be able to run anything they would on similar x86 devices. We’re running and fast catching up