- Mon Dec 29, 2014 5:31 pm
#5914
If I am correct, each function has a prefix i.e. ICACHE_FLASH_ATTR that determines where the function will be stored in flash. However if I am correct we do not have the source of the SDK files. Just the library files and a file (eagle.rom.addr.v6) giving the start adresses. Thus no link errors will be found. I have to admit that this is my personal guess work based on linking issues from my CP/M time (a very very long time ago) but it makes sense to me. I wouldn't even be surprised if one could steal some storage away from the range above 0x40000 for user program as currently with SKD0.9.2 the binary is something like 145K (of 256K). But I would have to try that out 
But even if it does work it mightnot with a next release.
But like I said before this is based on gut feeling and (probably) ancient linker experience (end 70-ties/beginning 80-ties) when I modified microsoft libraries (replacing buggy compiler modules). And the actual linker proces is still more-or less the same.
I just never have seen the sources of the actual SDK itself. Maybe they exist but I am not looking right
But even if it does work it mightnot with a next release.
But like I said before this is based on gut feeling and (probably) ancient linker experience (end 70-ties/beginning 80-ties) when I modified microsoft libraries (replacing buggy compiler modules). And the actual linker proces is still more-or less the same.
I just never have seen the sources of the actual SDK itself. Maybe they exist but I am not looking right
