How to implement this constraint in CVX?

I will assume that x and x_1 are both optimization variables. Depending on the values of A and b, this is not necessarily convex. if you can consolidate the RHS into the LHS and end up with a convex quadratic on the LHS, then it will be convex.