What does code pattern like .size X,.-X do?

Question

My question is about some popular but not well documented code I have found in ARM CORTEX M startup files like this STM32.

The 'pattern' is:

.size X,.-X

,where X is a symbol or label.

I have found this answer and understand how .size directive and dot special symbol work, but still complete line seems to do nothing to me. The result of the operation .-X isn't stored anywhere.

Could anybody explain what the line does?

Coldiron

The result is stored in the metadata automatically by the .size directive. – Pettifogger

The size calculation works exactly like $ - X in NASM: How does $ work in NASM, exactly?. But the .size directive emits ELF metadata with the function size, which is useful for shared libraries. – Sprinkle

Thank you for answers! What will happen if I do not put this code after the function? Will it have negative impact on optimization only or my code won't work as expected? – Coldiron

The effect of leaving it out is stuff like ARM Assembly Functions Appear With Size 0 in Symtab. I'm not sure if there are implications for correctness of static or dynamic linking. But it's just metadata, so if it works at all, there aren't any performance implications. – Sprinkle

If you miss it out it might affect how well the debugging of the assembler works for the reasons in the comment above. – Craquelure

Answer 1

That is placed at the end of function X, and the size of the function is the difference between the end of the function and the beginning. . is the current location so it's saying .size x is the difference between here and the label x.

Gypsophila