3.6 The short-term scheduler selects from the ready process the next process to run and gives it to the CPU. The long-term scheduler selects from the pool of processes that are waiting on disk and loads the selected process(es) into memory. These processes have not yet begun their execution. The medium-term scheduler takes processes that are currently in memory and selects those to be swapped out to disk. These processes will be swapped back in at a later point. This is done to improve process mix or because of memory requirements (i.e., overcommitted and needs to free up memory). A primary difference is in the frequency of their execution. The short-term scheduler must select a new process quite often. Long-term is used much less often since it handles placing jobs in the system and may wait a while for a job to finish before it admits another one.

3.7 In response to a clock interrupt, the OS saves the PC and user stack pointer of the currently executing process, and transfers control to the kernel clock interrupt handler, the clock interrupt handler saves the rest of the registers, as well as other machine state, such as the state of the floating point registers, in the process PCB. The OS invokes the scheduler to determine the next process to execute, the OS then retrieves the state of the next process from its PCB, and restores the registers. This restore operation takes the processor back to the state in which this process was previously interrupted, executing in user code with user mode privileges.

3.14 Synchronous and asynchronous communication: A benefit of synchronous communication is that it allows a rendezvous between the sender and receiver. A disadvantage of a blocking send is that a rendezvous may not be required and the message could be delivered asynchronously. As a result, message-passing systems often provide both forms of synchronization. Automatic and explicit buffering: Automatic buffering provides a queue with indefinite length, thus ensuring the sender will never have to block while waiting to copy a message. There are no specifications on how automatic buffering will be provided; one scheme may reserve sufficiently large memory where much of the memory is wasted. Explicit buffering specifies how large the buffer is. In this situation, the sender may be blocked while waiting for available space in the queue. However, it is less likely that memory will be wasted with explicit buffering.

Send by copy and send by reference: Send by copy does not allow the receiver to alter the state of the parameter; send by reference does allow it. A benefit of send by reference is that it allows the programmer to write a distributed version of a centralized application. Java’s RMI provides both; however, passing a parameter by reference requires declaring the parameter as a remote object as well.

Fixed-sized and variable-sized messages:
The implications of this are mostly related to buffering issues; with fixed-size messages, a buffer with a specific size can hold a known number of messages. The number of variable-sized messages that can be held by such a buffer is unknown. Consider how Windows 2000 handles this situation: with fixed-sized messages (anything < 256 bytes), the messages are copied from the address space of the sender to the address space of the receiving process. Larger messages (i.e. variable-sized messages) use shared memory to pass the message.

3.15 Yikes!

4.7 Any kind of sequential program is not a good candidate to be threaded. An example of this is a program that calculates an individual tax return. (2) Another example is a “shell” program such as the C-shell or Korn shell. Such a program must closely monitor its own working space such as open files, environment variables, and current working directory.

4.8 The user-level threads are known only within a given process. To context switch, we only need to save the thread-specific context: the program counter, CPU