Carlo Rovelli in one of his articles from 2019 (reference: https://arxiv.org/abs/1912.04702) argues that time travel into the past are thermodynamically impossible:
For instance, if we want to travel to the past and arrive to the past keeping our memory of events happened in the future, we need some device (like our brain) capable of memory. But memory is an irreversible phenomenon (we remember the past not the future) and, like all irreversible phenomena, is based solely on the only fundamental irreversible law: the second principle of thermodynamics $\frac{dS(\tau)}{d\tau}\ge 0$. Along a closed timelike loop (CTC) $\gamma$ the only possibility of having $\frac{dS(\tau)}{d\tau}\ge 0$ everywhere is having $\frac{dS(\tau)}{d\tau}= 0$. But this means that all the processes around γ are reversible, and therefore there can be no memory.
Where $\tau$ parametrizes the loop $\gamma$.
Since S is a state function, we must have $\Delta S = 0$ along a CTC...However, if the change in entropy of the system that runs through the CTC is zero, it does not imply that the process is reversible ... this is the thing I don't understand.
Perhaps Rovelli means that the device that runs through the CTC performs a reversible transformation and irreversibility is "discharged" into the external environment? I'm not sure about the latter explanation, any help is appreciated.
Addendum
If we consider a physical system along a CTC, then we can write: $\Delta S_{Universe}=\Delta S_{surrounding}+\Delta S_{system}$;
$\Delta S_{system}=0$ for any process along the CTC, but $\Delta S_{surrounding}\ne 0$ in general, thus $\Delta S_{Universe}\ne 0$.
So, how can we say that any process is reversible along the CTC?