Timeline for What is the relationship between system calls, message passing, and interrupts?
Current License: CC BY-SA 4.0
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 22, 2019 at 22:23 | comment | added | LawrenceC | Resets are lumped into the same category as interrupts because they follow the same basic logic: CPU receives a signal, CPU jumps to a location based on a pointer. Different types of interrupts are handled differently anyway - for example SMM saves different state than IRQ or NMI and even has its own return instruction, RSM. Also: A CPU reset isn't necessarily a full platform reset. State can be saved or discerned from I/O, devices, or RAM. For example, the Intel TCO watchdog timer has a some bits that a CPU can read after a reset to see if the watchdog caused the reset or something else. | |
| Jul 27, 2018 at 14:08 | history | edited | sakisk | CC BY-SA 4.0 |
improve formatting
|
| Nov 28, 2015 at 16:35 | comment | added | sakisk | @mezamorphic they refer to the question of the op | |
| Nov 28, 2015 at 16:19 | comment | added | intrigued_66 | You should list what "1" and "2" are- you refer to them but they are not listed anywhere. | |
| Aug 4, 2011 at 22:52 | vote | accept | Tim | ||
| Aug 3, 2011 at 13:21 | comment | added | psusi | @faif reset and NMI ( Non Maskable Interrupt ) are two different pins on the CPU. The latter causes an interrupt, meaning to save state and jump to handler. The former initializes the CPU ( intentionally discarding all state ) just as it does at power on ( reset is automatically asserted at power on ). Since part of the definition of an interrupt is that it saves state, reset fails to meet that definition. | |
| Aug 3, 2011 at 9:25 | comment | added | sakisk | @Tim I mean in a sense, yes. The simplest way to think of it is that traps can be generated by writing code in a programming language thus you are aware of them, while interrupts are generated by the hardware and are transparent to you. | |
| Aug 3, 2011 at 9:22 | comment | added | sakisk | @psusi I think that reset is an interrupt - non-maskable, but you're right that it doesn't need to save the state. | |
| Aug 3, 2011 at 2:50 | comment | added | psusi | A system reset is not an interrupt; it doesn't save state to resume after. | |
| Aug 2, 2011 at 23:24 | comment | added | Tim | Thanks! Regarding (3), do you mean in a sense no? What are the general purposes of hardware interrupt and of software interrupt? | |
| Aug 2, 2011 at 23:13 | comment | added | sakisk |
(1) Correct. (2) Well, system calls and message passing are two different techniques that can be used for achieving Interprocess Communication. But it's true that both are software-based, thus to get attention from the CPU you execute a trap (software interrupt). (3) In a sense, yes. An example of asking for a kernel service is executing a library procedure (which will be translated into 1 or more system calls) like creat for creating a new file. An example of an interrupt is the printer sending a signal to the CPU to inform it that printing is completed.
|
|
| Aug 2, 2011 at 22:35 | comment | added | Tim | Thanks! (1) In 1, do you mean the two ways of getting attention from CPUs are software interrupt (the program with system call example) and hardware interrupt (the reset key example)? (2) In 2, do you mean that system call and message passing are two ways of software interrupt? (3) Is it correct that the purpose of software interrupt is for process to request kernel services and resources, and the purpose of hardware interrupt is not? If yes, what is the purpose of hardware interrupt? | |
| Aug 2, 2011 at 21:58 | history | answered | sakisk | CC BY-SA 3.0 |