The most annoying part of development right now is the fact that decompressing the loader takes foreeeeever. We should investigate why exactly, the few options that come to mind right now:
Is this just slow because a kernel debug build is huge (and we consequently should just disable compression for development?) but why is compression not a bottleneck then?
Is the lz4_flex implementation just very slow and inefficient? Maybe missing RiscV SIMD support? Maybe just slow in QEMU?
Or are there settings we can tweak to make this faster?
The most annoying part of development right now is the fact that decompressing the loader takes foreeeeever. We should investigate why exactly, the few options that come to mind right now: