Common controls include no primary antibody or use of normal rabbit IgG in ChIP as a negative control. In most cases, if you are studying the same factor in different tissues or cell lines, one control is sufficient. More controls will be needed if you are studying multiple factors from the same source.
We have seen researchers use either a genomic DNA input control or a "mock" control with IgG. A mock sample is usually generated by the following process: after sonication, divide your sample into two equal parts and add your antibody to one part, while adding mock antibody (normal IgG) or pre-immune serum to the other. The goal is to ensure that you are not seeing sequences that stick non-specifically to the antibodies and that no other part of the process is influencing your results (a process control). In addition, it ensures that the recognition of your protein by the antibody you are using is required for enrichment of the target sequence. We have received feedback from several successful ChIP-Seq researchers who have told us that a mock control is the best, with the lowest background, but this does require more work than an input control.