A hardware vulnerability is an exploitable weakness in a computer system that enables attack through remote or physical access to system hardware.
Any means by which code can be introduced to a computer is inherently a hardware vulnerability. That means that when a user installs software, moves files such as CD/DVD ROMs or plugs in flash drives those items can all be thought of as hardware vulnerabilities, as can interfaces on the board by which the drives are connected. Securing physical access by locking any rooms, cabinets and cases housing computer equipment protects against this type of vulnerability.
Another type of hardware vulnerability is an unexpected flaw in operation that allows attackers to gain control of a system by elevating privileges or executing code. These vulnerabilities can sometimes be exploited remotely, rather than requiring physical access.
One such exploit, Rowhammer, works by repeatedly rewriting memory in the same addresses to allow retrieval of data from nearby address memory cells – even if the cells are protected. While this is not supposed to happen, it can and does due to hardware flaws that are hard to prevent. Google Project Zero researcher Mark Seaborn and reverse engineer Thomas Dullien detailed two proof-of-concept (POC) attacks exploiting Rowhammer. As they explained, repeatedly accessing a row of memory can cause bit flips in adjacent rows of some DRAM devices.
Hardware vulnerabilities are not generally exploited through random hacking attempts but more typically in targeted attacks of known high-value systems and organizations. For most users, traditional malware protection and a locked door are sufficient protection.