• course outline
• .o, .a, .so
• building

• make

• kernel mode
• system calls
• exceptions
  • events triggering
  • what the hardware must do
  • exception handlers

• threads and context switches
• process = threads + address space
• Unix signals
  • events triggering
  • signal handlers and implementation (start)

• exceptions and the quiz

• Unix signals

  • signal handlers and implementation (finish)

• user IDs/account abstraction
• POSIX permissions, briefly
• set-user-ID, sudo is not kernel mode
• privilege escalation

• virtual memory concept
  • address space and address sizes
• page tables
• storing in memory (start)

• page tables
  • small example (not in memory)
  • storing in memory (finish)

• page tables (finish)
  • as a tree
  • multi-level lookup example
  • permission bits

• page tables: loading data on demand
• processes, threads
• creating new processes
  • fork, exec*
  • copy-on-write
• preview: threads
• waiting for processes

• briefly, pipe and dup2
• file descriptors
• pipe

• caching idea
• (if time) connecting processors to memory
• memory hierarchy
• direct-mapped caches
  • valid bits, tags

• cache misses and C code
• set-associative caches
• (briefly) replacement policies

• write policies
• the TLB

• creating new threads

• thread use examples

• problem of synchronization
• atomicity, race conditions

• reordering
• locks and mutexes
• barriers

• interlude: reordering

• deadlock
  • examples of

(Reiss out of town, Campbell guest lectured)

Unfortunately, the audio on the lecture recording is noisy.

Access lecture recording here

Or access this version with volume boosted: browser mp4

Or if that doesn’t work/you prefer me doing the lecture:

• I have a recording of me practicing going through the same topics outside lecture: mp4; or
• last semester’s 4 April lecture starting at 24 minutes in covers close to the same material: link to schedule from last semester

• deadlock
  • definition
  • preventing/avoiding
• producer/consumer problem
• monitors (start)

• monitors (finish)
• false sharing

• briefly, transactions

• mailbox analogy
  • best-effort delivery
  • envelope versus contents
• layers
• connections atop mailbox model (start)

• connections atop mailbox model (finish)
• names and addresses
• example: a frame
  • layers (link, network, transport)
• IP
  • addresses (v4, v6)
  • routers
• port numbers and UDP

• mapping between addresses and names/other addresses:
  • MAC <-> IP mapping
  • DNS
  • NAT
• URLs

• DHCP

• cryptography (start)
  • symmetric encryption

• cryptography (con’t)
  • asymmetric encryption
  • digital signature
• replay attacks
• public key infrastructure (start)

• public key infrastructure (finish)
• extra cryptography bits
  • cryptographic hashes
  • key exchange

• TLS protocol outline

• review: single-cycle CPU
• pipelining idea
• pipelining for the CPU (start)

• pipelining for the CPU (con’t)
• hazards and hazard handlings
  • stalling / waiting
  • forwarding and forwarding logic
  • speculation: guess and check later (start)

• speculation: guess and check later (finish)

• pipeline exercises

• briefly, branch prediction ideas (start)
  • static prediction

Note: the first few minutes of screencapture for the lecture recording is rerecorded (and so might not quite be in sync)

• briefly, branch prediction ideas (start)
  • local history
• idea of multiple issue + out-of-order
• out-of-order pipeline
• register renaming
• instruction queues

• data flow model for out-of-order

• out-of-order limits

• side channels generally
• side channels in the cache

• meltdown:
  • speculation past a priv/protection fault
• spectre
  • speculation past a bounds chcek
  • using cache side-channels

text file from lecture

Scheduled time: 9am-12pm Thursday 14 December in McLeod auditorium