Operating Systems Mind Map

Every operating systems visualization on this site on one map. The branches are the main ideas. Each leaf is a page; the grey line under it is the one thing to remember about it. Click a branch to focus on it, click a page to open it.

Recall mode hides the page names. Read the hint, say the name, then click the "?" to check.

1. How the map is organized

Overview of the operating systems mind map: Boot and program start, Processes and the kernel, Interrupts and devices, Scheduling and deadlock, Virtual memory, Physical memory, Files and storage, Disks and RAID, Isolation and security. — The map at a glance: 9 branches, 55 pages.

Ask which resource the kernel is managing on that page, and you know the branch:

Getting started? Firmware, linker and loader work before main() runs: boot and program start.
A process and its boundary with the kernel? System calls, fork, signals and pipes are processes; limiting what a process may see or touch is isolation.
CPU time or memory? Who runs next is scheduling. Translating addresses is virtual memory; handing out real frames and bytes is physical memory.
A device? How the device gets the CPU's attention is interrupts; how data reaches a file on disk is files and storage; the block devices under the files, alone or in an array, are disks and RAID.

2. The branches side by side

Branch	Resource	Key structure	Typical failure
Boot and program start	the machine, a binary	ELF file, GOT/PLT	kernel panic, missing .so
Processes and the kernel	a process	task, fd table	zombie, broken pipe
Interrupts and devices	CPU attention	IDT, IRQ line	interrupt storm, dropped packets
Scheduling and deadlock	CPU time	run queue	starvation, priority inversion
Virtual memory	addresses	page table, TLB	SIGSEGV, thrashing
Physical memory	RAM frames, bytes	free lists, malloc chunks	fragmentation, OOM kill
Files and storage	disk blocks	inode, page cache	data lost without fsync
Disks and RAID	disk sectors, flash pages	request queue, stripe layout	bad sector, slow rebuild, write hole
Isolation and security	what a process may do	cred, namespaces, cgroups	EACCES, EPERM

3. Five facts to remember

A system call switches the CPU to kernel mode; the kernel does the work and returns to user mode.
Each process has its own virtual address space; the MMU translates every address and the TLB caches the result.
fork() is cheap because of copy-on-write: a page is copied only when someone writes to it.
write() usually only dirties the page cache; the data is safe on disk only after fsync().
A container is an ordinary process: namespaces control what it sees, cgroups control how much it uses.

See all operating systems pages in the index, or the other mind maps.