The decision boundary will be perpendicular to $w$.
If the weight vector is multiplied by -1 , then for the data points on the RHS, $w^T x$ will be less than 0, and on the LHS, it will be greater than or equal to zero.
Accordingly, LHS data points (and hence $x_1$) will be predicted as +1 .
And RHS data points (and hence $x_2$) will be predicted as -1 .