Physical layer security (PLS) is an essential way to provide security to users in wireless networks against active or passive attacks. The quantification of the provided security is measured through secrecy outage probability and outage probability. For cognitive vehicular networks (CVN), PLS can increase the amount of the quality of services in infotainment applications. Due to the broadcast nature of the wireless channels, it is challenging to achieve the perfect secrecy between the transmitting and receiving nodes. Furthermore, in CVN, when resources are being shared in an underlay manner, the need arises to find the highest and lowest achievable quantification of performance measures for poor fading scenarios. The knowledge of the maximum achievable performance helps to design and develop the anti eavesdropping solutions. In a particular direction, firstly, this paper finds the closed-form expressions of first-order secrecy metrics over multiple input multiple output 2x2 Nakagami-m channel for vehicular networks. The derived closed-form expressions are further utilized to find the upper and lower bounds of secrecy outage probability, and outage probability using the integral bound method for CVN. The derived bound provides the knowledge of the region over a valid signal-to-noise ratio for which the maximum and minimum achievable performance can exist. To restrict the amount of eavesdropping, the convex optimization-based approach is proposed. The optimization method helps secrecy metrics to remain in a specific region of the value. Therefore, the work provides insights over the classical approach and the information-theoretic approach. These works enhance the knowledge to prepare the anti-eavesdropping methods for connected cars. Moreover, the derived analytical expressions demonstrate the impact of the wireless medium on secrecy performance while the derived bound demonstrates the impact of the magnitude of wireless fading parameters.