What this book covers
Chapter 1, Where It All Starts From – The Virtual Filesystem, provides an introduction to the virtual filesystem (VFS) in the Linux kernel. The chapter will explain the pivotal role of the VFS in the I/O stack and provide a strong conceptual understanding of VFS as it is the starting point of an I/O request in Linux.
Chapter 2, Explaining the Data Structures in a VFS, introduces the various data structures utilized by the VFS in the kernel. The usage of structures by the kernel such as inodes, directory entries, and file objects to store file metadata, directories, and open files will be explained. Furthermore, the method by which the superblock structure enables the kernel to record filesystem characteristics will also be covered. Finally, the page caching mechanism in the kernel will be explained.
Chapter 3, Exploring the Actual Filesystems Under the VFS, introduces the concept of filesystems in Linux. One of the most popular block-based filesystems in Linux, the extended filesystem, will be explained. Additionally, important filesystem concepts such as journaling and copy-on-write will be discussed in detail. The chapter will also delve into the differences between file and block I/O and explore the network filesystem. Finally, the concept of userspace filesystems will be introduced.
Chapter 4, Understanding the Block Layer, Block Devices, and Data Structures, introduces the block layer in the kernel. The chapter will explain the concept of block devices and how they differ from character devices, and covers the major data structures in the block layer.
Chapter 5, Understanding the Block Layer, Multi-Queue, and Device Mapper, introduces the device mapper framework and the multi-queue block I/O queuing mechanism in the kernel. The chapter will explain how the multi-queue framework improves the performance of I/O operations on modern storage devices. Further, the role of the device mapper in laying the foundation for features such as LVM will be explained.
Chapter 6, Understanding I/O Handling and Scheduling in the Block Layer, discusses the different I/O handling mechanisms and I/O schedulers in the kernel. Operations such as merging, coalescing, and plugging are discussed in terms of how they are handled in the block layer. The different I/O schedulers supported in the Linux kernel are also explained in detail, along with the differences in their operational logic.
Chapter 7, The SCSI Subsystem, focuses on the SCSI subsystem in Linux and its multi-layered architecture. The chapter will explain the multilayered SCSI architecture, the SCSI device addressing mechanism, and the main data structures in the SCSI layer.
Chapter 8, Illustrating the Layout of Physical Media, discusses the different storage media available today, and the differences in their architecture. The traditional mechanical drives, solid-state drives, and the new NVMe interface are compared.
Chapter 9, Analyzing Physical Storage Performance, covers the performance analysis and characteristics of the storage subsystem. This chapter will present the different metrics that can be used to assess the performance of physical storage. Later on, the different tools and mechanisms that can be used to gauge storage performance are discussed.
Chapter 10, Analyzing Filesystems and the Block Layer, focuses on the techniques that can be used to analyze the performance of the block layer and filesystems. This chapter will explain the different types of filesystem I/O and the factors that can affect the I/O request of an application. Later on, the different tools and tracing mechanisms, such as the Berkeley Packet Filter compiler collection, are looked at to identify potential bottlenecks at each layer.
Chapter 11, Tuning the I/O Stack, discusses some recommended practices for tuning the underlying storage layer for application needs. The different options that can be used to tune the performance at each layer are discussed.
