prediction <- predict(cv.fit, testData$mpg, s="lambda.1se")
It seems that testData$mpg is a vector, model should use the whole testdata set to predict instead of the single mpg values.
In your case, it should be something like
testdata <- as.matrix(data.frame(cylinderDummy[,2:ncol(cylinderDummy)], testData$displacement,
testData$horsepower, testData$weight, testData$acceleration,
originDummy[,2:ncol(originDummy)]))
prediction <- predict(cv.fit, testData, s="lambda.1se")