Operating Systems · Week 3

Mutexes, spinlocks and the cost of waiting

01 / 13Present00:00

Operating Systems • Week 3

01 / 13

Mutexes, spinlocks and the cost of waiting

A race condition is a bug that only shows up when you're not looking.

StackShade deck

Predict → explain → retrieve

COURSE CONTENTS

Mutexes, spinlocks and the cost of waiting

01OS Foundations & System CallsWeek 1 · 4 topics0/4

What an OS actually does: abstraction + arbitration12 min User mode vs kernel mode, rings, and the syscall path14 min Interactive: trace a read() from C code to disk12 min Interrupts, traps and the timer that keeps processes honest12 min

02Processes & ThreadsWeek 1–2 · 5 topics0/5

Process anatomy: text, data, heap, stack, PCB14 min fork(), exec(), wait(): how Unix births processes14 min Threads: shared address space, private stacks/registers12 min Interactive: context switch — what gets saved where12 min IPC: pipes, shared memory, message queues, sockets14 min

03CPU SchedulingWeek 2 · 5 topics0/5

FCFS, SJF, and the convoy effect12 min Round Robin: picking the quantum sweet spot12 min MLFQ: how real schedulers approximate SJF without knowing the future14 min Interactive: scheduling simulator — race the algorithms15 min Compute avg turnaround for 3 policies (by hand)25 min

04Concurrency & SynchronizationWeek 3 · 5 topics0/5

Race conditions and the critical-section requirements14 min Mutexes, spinlocks and the cost of waiting12 min Semaphores, condition variables, monitors14 min Interactive: producer-consumer — watch the lost wakeup12 min Fix 3 buggy concurrent programs (find the race)35 min

05DeadlocksWeek 3–4 · 4 topics0/4

Coffman conditions: mutual exclusion, hold-and-wait, no preemption, circular wait14 min Prevention vs avoidance: the Banker's Algorithm14 min Detection & recovery in real systems10 min Banker's Algorithm safety check (by hand)25 min

06Memory Management & Virtual MemoryWeek 4–5 · 5 topics0/5

Address spaces, base/bounds, segmentation (and its flaws)14 min Paging: pages, frames, page tables14 min TLB and multi-level page tables — making translation fast and small14 min Interactive: translate a virtual address step by step12 min Copy-on-write, memory-mapped files, swapping12 min

07Page Replacement & ThrashingWeek 5 · 4 topics0/4

Optimal, FIFO (and Belady's anomaly), LRU, clock14 min Working sets, locality and thrashing12 min Interactive: run reference strings through FIFO vs LRU12 min Count page faults for 3 policies (by hand)25 min

08File Systems, I/O & VirtualizationWeek 6 · 5 topics0/5

Inodes, directories, links — how a path becomes bytes14 min Journaling and crash consistency12 min Disks, SSDs, DMA and I/O scheduling12 min Virtual machines vs containers: what's actually shared14 min Capstone: trace a file write end-to-end with strace30 min