What is overflow_usub?

Question

Linus Torvalds has recently made it to mainstream news with a rant over a pull request. This pull request included a function, overflow_usub, which is apparently non-standard and uses some kind of compiler magic. As a result of the widespread reporting of this rant, it is near-impossible to find any useful information about this function. My question is: what is overflow_usub, when should it be used and what kind of compiler magic does it require?

Answer 1

The function overflow_usub is defined as:

static inline bool overflow_usub(unsigned int a, unsigned int b, unsigned int *res){   *res = a - b;   return *res > a ? true : false; } 

It will check for integer overflows in subtraction and doesn't involve any compiler magic. It's usually a fallback, if the compiler has no __builtin_usub_overflow.