Glossary

Relevant source files

This glossary defines technical terms, acronyms, and core data structures used throughout the Linux kernel. It serves as a high-level reference for the codebase at https It groups terminology into core kernel concepts and specialized subsystem terms.

Overview of Terminology

The Linux kernel codebase uses a mix of standard operating system nomenclature and kernel-specific abstractions. These terms are generally categorized into core infrastructure (processes, memory, scheduling) and subsystem-specific implementations (virtualization, storage, networking).

Core Architecture Concepts

The following diagram illustrates relationships between high-level natural language concepts and their primary data structure representations in the kernel core.

Kernel Core Entity Mapping

Sources: include/linux/sched.h1-100 include/linux/mm.h1-100 kernel/sched/core.c11-41

12.1 Core Kernel Terms

This section defines fundamental terms related to kernel execution and memory management. Each term includes key structs or functions for reference. For full technical depth, see Core Kernel Terms (process, memory, scheduling, tracing).

Term	Definition	Key Code Pointers
task_struct	The primary descriptor for a process or thread, holding its state, context, and scheduling info.	`struct task_struct` include/linux/sched.h748-1500
mm_struct	Represents the virtual address space of a process, managing VMAs and page tables.	`struct mm_struct` include/linux/mm_types.h615-815
VMA (Virtual Memory Area)	Represents a contiguous virtual memory range with uniform permissions.	`struct vm_area_struct` include/linux/mm_types.h660-750
Folio / struct page	Physical memory units; `folio` represents a non-tail page or a group of pages.	`struct folio` include/linux/mm_types.h294-400
GFP (Get Free Page)	Flags specifying allocation constraints (e.g., `GFP_KERNEL`, `GFP_ATOMIC`).	GFP flags include/linux/gfp_types.h1-100
RCU (Read-Copy-Update)	Synchronization mechanism allowing lockless reads concurrently with updates.	`rcu_read_lock()` include/linux/rcupdate.h1-100
Maple Tree	Range-based B-tree used for VMA management, replacing the red-black tree.	`struct maple_tree` include/linux/maple_tree.h1-50
sched_ext (SCX)	Extensible scheduler class that allows scheduling policies to be implemented in BPF.	`struct sched_ext_ops` kernel/sched/ext.h1-100
SLUB allocator	The default slab allocator, using per-CPU sheaves and NUMA-node barns for object caching.	`kmem_cache` mm/slub.c1-160
workqueue	Mechanism for offloading work to be executed later in process context.	`struct workqueue_struct` kernel/workqueue.c1-100
tracefs/ftrace	The interface and engine for function tracing and dynamic events.	kernel/trace/trace.c1-100
perf_event	Subsystem for hardware/software performance counters and sampling.	`struct perf_event` include/linux/perf_event.h1-100

Sources: include/linux/mm.h294-800 kernel/sched/core.c11-41 mm/slub.c1-160 kernel/sched/sched.h1-100

12.2 Subsystem Terms

Kernel subsystems implement specialized functionality on top of the core. For more details, see Subsystem Terms (KVM, BPF, io_uring, filesystems, drivers).

Filesystem and Storage Subsystems

The Virtual File System (VFS) layer abstracts various filesystems behind a common interface.

VFS Object Relationship

Sources: include/linux/fs.h1-1000 fs/dcache.c1-100

Virtualization (KVM) and Confidential Computing

KVM provides infrastructure for hardware-accelerated virtualization.

KVM Architecture Mapping

Sources: arch/x86/kvm/x86.c143-150 arch/x86/kvm/vmx/vmx.c1-100 arch/x86/kvm/svm/svm.c1-100 arch/x86/kvm/svm/sev.c55-65

Term	Definition	Key Code Pointers
KVM_RUN	The primary ioctl to start/resume guest execution.	`KVM_RUN` virt/kvm/kvm_main.c4100-4200
VMExit	Transition from guest mode to host mode (hypervisor).	`handle_exit` arch/x86/kvm/vmx/vmx.c6200-6300
TDP MMU	MMU optimized for Two-Dimensional Paging (EPT/NPT).	arch/x86/kvm/mmu/tdp_mmu.c1-50
SEV / TDX	Hardware features for memory encryption and guest isolation.	arch/x86/kvm/svm/sev.c1-100 arch/x86/kvm/vmx/tdx.c1-100
Verifier	BPF component that ensures program safety (no loops, valid memory access).	`bpf_check()` kernel/bpf/verifier.c57-170
io_uring	Asynchronous I/O interface using shared submission/completion rings.	`struct io_ring_ctx` io_uring/io_uring.c1-100
zcrx	Zero-copy receive for io_uring, utilizing page pools and ifqs.	`struct io_zcrx_ifq` io_uring/zcrx.c35-56

Sources: kernel/bpf/verifier.c1-100 io_uring/io_uring.c1-100 arch/x86/kvm/x86.c1-200 io_uring/zcrx.c1-100

For deep dives on these concepts and others across the kernel, please consult the child pages:

Sources: include/linux/sched.h1-100 include/linux/mm.h1-100 kernel/sched/core.c1-100 mm/slub.c1-160 arch/x86/kvm/x86.c1-200 include/linux/fs.h1-1000 kernel/bpf/verifier.c1-100 io_uring/io_uring.c1-100