[userbinator]

    -Os

Then if any functions are a bottleneck, compile them with function-specific O2 or O3.

[nly]

-Os doesn't always produce smaller code than -O2

[joveian]

It also unfortunately breaks a lot of things and exposes more gcc bugs than the commonly used options. I wish "make the output code as small as possible" was the usual thing compilers tried to do.