With robust and ubiquitous network connectivity becoming increasingly important for cyber vehicles, roadside infrastructure plays a critical role in the cyber transportation system. Data sent to roadside units (RSUs) however, can leave traces of the vehicle that created them. This data can leak sensitive information such as driving patterns and driving history. In this paper we consider the vehicle location privacy problem in regards to communication with RSUs. We propose to use delay tolerant networks (DTNs) with vehicles as nodes to mask the location of the vehicle through a pass and run protocol. Using vehicle-to-vehicle communication, packets are sent through the DTN until a certain condition is met, at which point they will be sent to the nearest RSU. The vehicle that created the packet will pass the packet and run, thus the packet and the vehicle will travel independently to effectively hide the location of the vehicle. We give an abstract notion of perfect privacy and define two quantitative measurements for location privacy: the distance metric is based on the distance between vehicle's real locations and the location when the data packets are received, while the direction metric is the angle between vehicle's driving direction and the moving direction of received packets. Simulation based on real life taxi information from the public domain demonstrates that the proposed pass and run protocol can effectively preserve vehicle's location privacy.