The challenging and more urgent work is to replace lower-level proprietary components on these systems with FLOSS alternatives,
I don’t think that device drivers are really the first problem the free software movement needs to tackle. I believe that the most important issue is to have good applications, with superb usability and that are innovative in order to attract users, fast. Some of the tactics used in the GNU project will need to be adapted to the speed of mobile, while others are not applicable.
Stallman’s project started in a time when PCs were slowly becoming relevant in society. It took almost 10 years before they were cheap enough to be in the bedrooms of young, smart programmers for them to easily contribute to the project. GNU also started developing applications first, and it took almost 10 years to start working on kernel and device drivers. The early adopters of GNU were highly skilled users, in a world with few computers with a clearly winning platform (the standard/commoditized platform IBM/Intel x86). Stallman and the whole free software movement had a lot of time to develop a nice free-as-in-freedom operating system and applications on standardized hardware. They also had the Unix design to follow: how the system had to look was pretty clear, it only had to be ‘better’.
Compare those first ten years and the quantity of computers in the ’80s/’90s to today’s speed and the quantity of mobile devices in everybody’s pocket (not just in developed countries), without a clear plan to follow(like Unix was for GNU): the game is radically more challenging. Take Google’s G1 as an example: it’s only one year old but its operating system version is obsolete (and customers are complaining). With users changing phone every 18 months in the US, the lifecycle of a free driver is too short to justify the effort.
On the other hand there are many applications that need to be liberated, like social applications that respect freedom in the cloud, mobile email client that don’t suck, mobile music players with stores that are not defective-by-design. And many more need still need to be imagined. Developer’s focus should be on what appears in freedom-giving mobile applications markets: we made the application market concept popular (apt-get repository anyone?), now we need to move to mobile and to fill them with good and free applications first. Device drivers can come at later stage, eventually after hardware manufacturers will have battled each other to the death and one winner will emerge (like it happened with x86).