Program does not have enough pages - memory swapped
Continues to not have enough pages
Page faults are used to get pages from swap.
This swap makes the length it takes to complete the process exponentially longer.
Spends all time swapping memor
Geometrically close to the CPU.
3-4 clock cycle rw (1ns).
You can send in 64 bytes of data to the cache at a time.
Temporal Locality and Spatial Locality let you create code that uses the cache reliably.
Associated with 64 byte sized cache
Journaling File System
Records each update to the fs as a transaction.
Every operation is written to a journal that keeps track of all the edits that are made.
E.g. Transaction 110 - Write Hello.txt (Byte) 453: Hello.
File system is managed separately.
Groups of 4-8 KB of contiguous memory.
Physical memory pages are also known as frames
Page Table (Virtual Memory)
Resides in physical memory.
Each process maintains its own page table. i.e. unique to each process
Essentially a big hash table (used
Avoiding Buffer Overflow
Put secret number between frame pointer/return address and code.
Before returning, check the secret number and if it is different than the initial, then it is not
Breach of Confidentia
Access Lists and Groups
Three classes of users
Access Control Lists
More expressive than Access Lists
Specify different rights per user or group.
In Linux there is only 1 group, in Windows, there are multiple group
Harddisk, SSD, Flash Drive.
Read and write in 1KB blocks.
Create logical blocks using dd.
Keyboard, Network Card
Always chronological reading of bytes.
File systems are used in Block Devices.
File System Prope
No code writing required.
Code changing may happen.
How is it organized? What is a kernel? What are userland applications (dev tools, libc, system
tools)? Check first lecture.
Contains User Interface.
Multi-Level Page Tables
Split into 4 parts (if 4GB - 1GB each).
Requires first 2 bits to index into any of the 4 parts.
Split into pointers to 4KB physical frames.
Requires next 18 bits to unique address every page in 1GB space.