According to a web page the FAT parent *was* invalidated in 2004. IANAL but surely claiming violation of an invalid patent is useless. I doubt TomTom's kernels include FAT support anyway.
I am one of those that thinks algorithms and data structures are mathematics, so never patentable in any circumstances. Did Fredman and Tarjan miss is a trick by not patenting Fibonacci heaps? Nobody knows a priority heap structure that is asymptopically better.
The USPTO *should* be able to uncover relevant journal articles and read them. Given some of the patents granted I have my doubts about whether they can.