RFC: Using per function sections

[Joel Sherrill]

Hi

After Anthony Green commented how much code space was
saved by using per function sections, I have been doing some
experimenting on a few targets.  My impression is:

+ Some tests drop > 50% in size
+ Most drop 20-30%
+ Minimum drop is ~10%
+ Even a well tuned size constrained BSP like the lpc23xx_tli800
    saw 6-9% minimum improvements. At least one test fit into
    memory which did not before hand.

PowerPC and ARM appear to work great. SPARC/sis resulted in 80
byte .text sections.

I would like to know if there is general interest in using these
options across all BSPs where it works.  And investigating those
which do not.

-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill at OARcorp.com        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
Support Available                (256) 722-9985