With the continuous advancement of computer hardware architecture, the performance of storage devices has significantly improved. The emergence of new storage devices (such as Optane SSDs) has greatly reduced the overhead introduced by hardware devices in the I/O path. However, other components within the I/O stack still present notable performance bottlenecks, limiting the full potential of storage device performance. This paper focuses on optimizing the I/O stack of file system-based software read/write system calls, addressing various bottlenecks in the existing I/O stack, such as context switch overhead, physical page allocation, cache page lookup, dirty page marking, I/O aggregation, and scheduling. The paper provides a detailed overview and analysis of the existing optimization solutions. Additionally, the paper summarizes the advantages and disadvantages of these optimization techniques and proposes future research directions, aiming to provide theoretical support and practical guidance for optimizing the I/O path in high-performance storage device environments.