Sonar Acoustic feedback remains a critical limitation in sound reinforcement and communication systems, particularly in environments where microphones and loudspeakers operate close proximity. Traditional suppression methods such as notch filtering, phase shifting, and frequency shifting provide partial relief but often introduce latency, tonal coloration, and reduced audio quality. Recent work has increasingly explored adaptive approaches and hardware acceleration to address these limitations. In parallel, Field-Programmable Gate Arrays (FPGAs) have emerged as attractive platforms for real-time audio processing due to their parallelism, deterministic timing, and flexibility. This paper presents a comprehensive review of algorithms and implementation strategies for acoustic feedback suppression with emphasis on FPGA-based realization. Traditional and adaptive algorithms, including LMS-based feedback cancellation and transform-domain methods, are examined in terms of performance trade-offs, hardware cost, and suitability for real-time deployment. The review also analyzes practical challenges such as clock synchronization, resource utilization, precision management, and system scalability. A synthesis of the literature reveals key research gaps, including the scarcity of complete FPGA implementations, limited consideration of multi-microphone systems, minimal real-time testing, and lack of comparative benchmarking across algorithms on common hardware platforms. To address these gaps, the paper proposes a review-driven design perspective focusing on adaptive FPGA architectures, resource-optimized LMS variants, multi-path modeling, and systematic real-world validation. The findings demonstrate that FPGAs hold significant potential as enabling platforms for next-generation, low-latency acoustic feedback suppression systems capable of moving from laboratory prototypes to practical deployment.