This depends on your (top secret?) toolchain: Split flash area in linker/loader script into several parts, one 'default' part where the usual .text sections etc. go, and one separate part for each block (maybe individual flash pages, or even smaller blocks) which must be kept independent of each other. Then, in the sources, prefix each of the special functions with the proper pragma/attribute to indicate in which section it should go.
This will ensure each of your special functions will be located *somewhere* in one of the prescribed blocks. If you want precise address for each of these functions, just use a separate block for each one. Note that this gets slightly cumbersome, as you won't know the required size for each section in advance, so some experiments are necessary (and the size will change with each change of compiler version and settings).
Regarding the syntax for your particular toochain: Look into the Cube samplex, e.g. for the STM32F746-Disco where there are two separate flash sections end even several RAM sections.
