On http://fun-tech.se/stm32/OlimexBlinky/ld_float.php you mention a workaround for ld\'s inability to link the float library functions. It appears that this is caused by the need for multi-pass symbol resolving. The workaround you mention does work but it\'s unfortunately something I\'d consider a hack. A cleaner approach to solving the issue is to make ld repeatedly resolve the symbols of the object files and the required gcc library:
Thanks to Soeren Apel:
On http://fun-tech.se/stm32/OlimexBlinky/ld_float.php you mention a workaround for ld\'s inability to link the float library functions. It appears that this is caused by the need for multi-pass symbol resolving. The workaround you mention does work but it\'s unfortunately something I\'d consider a hack. A cleaner approach to solving the issue is to make ld repeatedly resolve the symbols of the object files and the required gcc library:
More or less change
into