Better handling of integer wraparound in the kernel

INT_MAX - b overflows for all negative b, not just INT_MIN. The ideal way to do this is with the compiler intrinsics or stdchkdint.h (C23). If you insist on doing it with regular arithmetic, then a + b never overflows for positive a and negative b (or vice-versa), nor will overflow happen if either operand is zero, so you can write some godawful nested ternary expression to check for the four possible cases (both positive, both negative, signs differ, either operand is zero) and hope the compiler is smart enough to understand what you are doing.

But don't actually do that. Just use the intrinsic.