[LLVMdev] Floating point atomic load and add

[Tyler Denniston]

As a followup: could I do a "strong" cmpxchg and leave out the loop?  
(According to  
http://en.cppreference.com/w/c/atomic/atomic_compare_exchange that is  
preferable when possible).

Here is (my understanding of) the equivalent LLVM:

void emitAtomicFPAdd(Value *ptr, Value *incr) {
   LoadInst *oldval = builder->CreateLoad(ptr);
   Value *newval = builder->CreateAdd(oldval, incr);
   Instruction *cmpxchg =
     builder->CreateAtomicCmpXchg(ptr, oldval, newval,
				 AtomicOrdering::SequentiallyConsistent,
				 AtomicOrdering::SequentiallyConsistent);
}

Quoting Tim Northover <t.p.northover at gmail.com>:

>>   } while (__c11_atomic_compare_exchange_weak(
>>       addr, &oldval, newval, memory_order_seq_cst, memory_order_relaxed));
>
> Actually, I think this condition is inverted. Should be "while
> (!_c11...". Sorry about that.
>
> Tim.