SIMD-based blitting using ARM Neon and Intel AVX

[nfeske]

At the moment, we use SIMD instructions for 2D-pixel copying only on the x86 platform. I'd like to foster the use of SIMD on ARMv8 as well, and also leverage more modern SIMD variants (AVX) on x86_64.

[nfeske]

Implemented for ARM Neon and SSE on my simd branch https://github.com/nfeske/genode/commits/simd/.

[nfeske]

My updated simd branch contains quite a few changes made after benchmarking on the PinePhone. The non-SIMD (slow) variant reads sequentially now, instead of writing sequentially. This is useful because the source is usually cached memory whereas the destination is uncached in most display drivers.

The Neon variant has been improved for cache locality and - most importantly - cache prefetching in the rotation case. On the PinePhone, the rotation adds about 40% overhead compared the non-rotated back2front copy. A complete screen update (1440x720) takes about 5.4 ms w/o rotation, and 7.7 ms with rotation.