Expression of the double-sex gene is rather unusual, so it is not surprising that you are having difficulty understand mutations in it. The simplified diagram below (adapted from Gilbert’s Developmental Biolgy — available on-line) should help clarify the matter.
The key concept is that although the double-sex gene (dsx) is expressed in both male and female flies, the gene products (DsxF add DsxM) are different in the two cases. This is achieved by alternative splicing of the RNA transcript, controlled ultimately by X and Y chromosome dosage. (If you are unfamiliar with alternative splicing, see e.g. this section of [Molecular Biology of the Cell (http://www.ncbi.nlm.nih.gov/books/NBK21558/).)
The second important concept (the details of which, I admit, I am less familiar with) is that sex is determined by both activating genes specific for the particular sex and inactivating genes for the opposite sex. Thus, the diagram shows that DsxF both activates expression of genes for female sex differentiation and represses expresses of genes for male sex differentiation. Likewise DsxM activates expression of genes for male sex differentiation and represses expresses of genes for female sex differentiation.
Thus, according to this schema, if dsx is completely knocked out there will be consequences in both male and female. In the female, in the absence of DsxF, the genes for female sex differentiation will not be activated, but those for male sex differentiation will not be repressed. This presumably results in the intersex phenotype. Likewise, in the male, the absence of DsxM prevents the expression of the genes for male sex differentiation, but does not repress those for female sex differentiation. Again, I presume this results in the intersex phenotype.