be falign=1. Saves a little over 3k on an x86 "make defconfig" build.
# gcc-4.0 and older seem to suffer from these
$(if $(call cc_le,4,0),\
$(call check_gcc,OPTIMIZATIONS,-mpreferred-stack-boundary=2,)\
- $(call check_gcc,OPTIMIZATIONS,-falign-functions=0 -falign-jumps=0 -falign-loops=0,\
+ $(call check_gcc,OPTIMIZATIONS,-falign-functions=1 -falign-jumps=1 -falign-loops=1,\
-malign-functions=0 -malign-jumps=0 -malign-loops=0))
# gcc-4.1 and beyond seem to benefit from these