OFDM (Orthogonal Frequency Division Multiplexing) is a promising technology for future broadband wireless communications due to its capability in combating multipath fading. Carrier frequency offset (CFO), which can induce the loss of orthogonality among OFDM sub-carriers and significant performance degradation, needs to be estimated and compensated for. In this paper, we propose a general CFO estimator based on the maximum likelihood (ML) estimation criterion. It will be shown that with the same estimator architecture, CFO can be obtained using either training OFDM symbols, pilot tones, null sub-carriers, or a combination of them. Furthermore, by taking advantage of the channel side information, the performance of CFO estimation can be significantly improved. To demonstrate the capability of the proposed CFO estimator, we will consider an OFDM system using the signal structure of the IEEE WLAN standard 802.11a. Compared with previous work using null sub-carriers alone, it will be shown that by taking advantage of the pilot tones, null sub-carriers, and the channel statistics, the performance of CFO estimation can be improved by about 2dB.