coverity issue in generated CUDA kernels

[raffenet]

Coverity's static analysis does not like this pattern in the CUDA generated code. From yaksuri_cudai_kernel_pack_SUM_hvector_hindexed_int16_t.

[...]
    uintptr_t x3;
    for (intptr_t i = 0; i < md->u.hvector.child->u.hindexed.count; i++) {
            uintptr_t in_elems = md->u.hvector.child->u.hindexed.array_of_blocklengths[i] *
                                 md->u.hvector.child->u.hindexed.child->num_elements;
            if (res < in_elems) {
                    x3 = i;
                    res %= in_elems;
                    inner_elements = md->u.hvector.child->u.hindexed.child->num_elements;
                    break;
            } else {
                    res -= in_elems;
            }
    }
[...]

If the loop exist without the if (res < in_elems) branch ever being taken, then x3 will be uninitialized when it is used later on in the function.

1. What is uintptr_t res representing in this context? It could benefit from a better name.
2. How can we modify the loop to guarantee static analysis sees x3 as having a valid value?