Abstract: Wireless communication is vulnerable to malicious jamming and eavesdropping attacks due to the broadcast nature of wireless channels. Extremely large-scale reconfigurable intelligent surface (XL-RIS) demonstrates its abilities to enhance the physical layer security and compensate for the severe path loss. Due to these advantages of XL-RIS, this paper investigates an XL-RIS empowered near-field physical-layer-security communication system against jamming and eavesdropping attacks with the help of artificial noise (AN). To maximize the secrecy capacity, we propose an alternating-optimization (AO)-based algorithm to jointly optimize the beamformers at the base station (BS) and the reflection coefficient matrix at the XL-RIS, subject to the BS's maximum transmit power and the XL-RIS's unit-modulus constraints. For the beamforming and AN design at the BS, auxiliary variables are introduced to reformulate the sub-problem into a more tractable problem, which is solved by the proposed successive-convex-approximation based algorithm. For the reflection coefficient matrix design at the XL-RIS, a manifold-optimization based solving algorithm is proposed to address the challenge of large-scale variables and unit-modulus constraints. Numerical results show that (1) XL-RIS can ensure secure communication even if the eavesdropper is located at the same direction as the legitimate user and closer to the XL-RIS; (2) the proposed algorithm improves the secrecy capacity compared with the benchmark schemes.