Short answer; Almost NO!
Can I have it by keep hashing the second one?
There were great answers for the question; Cycles in SHA256
In short, if we model SHA256 as a uniform random function then the probability of element being on the cycle is
$$\frac{1}{\sqrt{\hspace{.03 in}2\hspace{-0.05 in}\cdot \hspace{-0.04 in}\pi} \cdot 2^{127}}$$
The average cycle length with expected value for SHA256 is $$2^{127} \sqrt{2\pi}$$
Simply consider the first hash $\operatorname{SHA256}(m)$ as the starting point;
- As one can see that being on a cycle has a very low probability and you almost find none in a cycle since the probability is $\frac{1}{\sqrt{2\cdot \pi} \cdot 2^{127}}$.
- Even it is on a cycle, you almost certainly can not calculate the cycle to find the pre-image since the average cycle length is $2^{127} \sqrt{2\pi}$.
These are from the result of Bernard Harris's magnificent work; Probability Distributions Related to Random Mappings in 1960.
Also, this exists in the Handbook of Applied Cryptography - Fact 2.37
- tail lenght = $\sqrt{\pi n /8}$
- cycle lenght = $\sqrt{\pi n /8}$
- rho-lenght = $\sqrt{\pi n /2}$