This question is a bit esoteric in nature but it was mentioned recently and I thought it might be an interesting post.
The Cortex M3 supports bitbanding...previous ARM architectures (ARM9) did not.
The question is generally speaking how transparent is the bitbanding feature being taken advantage of by the compiler.
Obviously the answers are "it depends on" coding style, amount of access to registers/memory locations supported by bitbanding etc.
Q1: Does the TI (or any other compiler for Cortex) M3 know how to take advantage of Bitbanding?
Q2: if my legacy code was optimized for ARM9, then when I recompile, is there a "trick' to get the Bitbanding optimization for M3?
Q3: Although unlikely, the scenario could also work the opposite direction...C code targeted for M3 would be recompiled for ARM9...what would be
the impact to code efficiency?
