Separate thread stacks make one thread's stack memory inaccessible to every other thread in the process.

What memory do threads within one process share?

• One address space, including globals and heap data
• Separate address spaces, each holding copies of all globals
• Only a disk directory, with no common memory addresses

Why does creating a thread usually need less address-space setup than creating a process?

• A thread never needs any stack storage or saved execution state
• A thread usually reuses the process's existing memory mappings
• A thread automatically runs on a separate, faster physical CPU

Two OS-scheduled threads share a process on one CPU. A waits for disk input, while B is runnable and has independent computation to do. What can happen?

• Another runnable thread can compute during the disk wait
• Both threads must stop until every outstanding disk request ends
• The waiting thread automatically releases all its memory mappings

A uses local variable a, B uses local variable b, and both update global total. Neither local variable is passed to the other thread. Which variable is explicitly accessed by both?

• a, the local value in thread A
• b, the local value in thread B
• total, the global value

Which part of the setup can a switch between two threads of one process keep unchanged?

• Keep the same saved program counter for both threads
• Skip saving or restoring any register state at all
• Keep the same process address-space mappings

A shared pair starts left = 1, right = 1. A writer changes both to 2. A reader needs matching values. Which coordination is needed?

• Coordinate the reader and writer using one shared rule
• Coordinate only the writer while the reader ignores its updates
• Give each access a separate rule and let the reader run at any time

In a non-atomic teaching trace, total starts at 4. A and B each read 4, add 1 locally, then store their result. A stores before B. What is the final total?

• 4
• 5
• 6