[FFmpeg-devel] [PATCH 0/3] [RFC] Optimizing XTEA

[Michael Niedermayer]

This patchset is an alternative to the optimization i just pushed
The problem with the patchset here is it breaks ABI due to the xtea.c/h 
ABI having a hardcoded and public context size.
The difference in terms of speed on a i7 between this patchset and
the pushed code is about 1%.

We could introduce a new XTEA ABI/API that doesnt have the problems
of xtea.h and deprecate xtea.c/h but iam uncertain this is worth it
for 1% ...