I know each thread has its own stack.
And my textbook says:
Suppose that a particular semaphore implementation uses a LIFO stack of threads for each semaphore. When a thread blocks on a semaphore in a P operation, its ID is pushed onto the stack. Similarly, the V operation pops the top thread ID from the stack and restarts that thread
what I don't understand is. since each thread uses its own stack to store its thread ID, and obviously threads don't share stack with others, so what does LIFO stack of threads mean? because a thread only put its thread onto the stack which won't contain another thread's id, why use LIFO term here, isn't that LIFO only be meaningful if we can push multiple items on stack?
Your OS needs to know which all threads are waiting for each semaphore. To maintain that list, OS can use a LIFO data structure i.e. stack for each semaphore object.
Do not confuse it with stack of each thread.
what I don't understand is. since each thread uses its own stack to store its thread ID, and obviously threads don't share stack with others, so what does LIFO stack of threads mean?
Threads can shared their stacks with others if they want. Each thread having its own stack is no impediment to one thread accessing another thread's stack.
because a thread only put its thread onto the stack which won't contain another thread's id, why use LIFO term here, isn't that LIFO only be meaningful if we can push multiple items on stack?
The fact that each thread has its own stack doesn't prevent others from existing for other purposes. That each thread has its own stack has no effect on what a sempahore can do or not do with a stack.
Think about apples. You can have them. You can share them. You can eat them.
Now, say there are ten children and each has their own apple. Does that mean there cannot exist an eleventh apple that's shared? Does this prevent them from doing other things with other apples? No, it does not.
So, yes, each thread has its own stack. And they can also share those stacks if they want.
But also, a semaphore can have a stack. And it can do anything it wants with that stack. This has nothing whatsoever to do with any other stacks that threads might have.
Related
I am an absolute beginner in operating systems. So please, do not mind if the question appears too naive or basic.
From what I've read, each process has its own Kernel stack and User stack. So does each thread. Threads of a process share the same address space. They also share the code and data segment, but not the stack.
But how is this possible? There is only one stack pointer in a CPU, so how can each thread have its own stack?
And what is the difference b/w stack and stack frame? From what I've read, there is only one stack and frames are pushed on it. Again, it is a physical stack? Do these stack exist in the virtual memory? Can someone please clear my concepts? I am confused and cannot move forward.
From what I've read, each process has its own Kernel stack and User stack. So does each thread.
Each thread has its own kernel and user stack. Processes may contain any number of stacks -- at least one for each of their threads, possibly more.
Threads of a process share the same address space. They also share the code and data segment, but not the stack. But how is this possible?
Because the term "share" is being used in two different ways.
My wife and I both jointly own two cars, so in that sense, we share two cars. But I have one car that only I use and she has one car that only she uses. In that sense, we each have our own car.
Similarly, a process with two threads has two stacks that are shared. One is for each thread. So each thread has its own stack, though they can access each other's stacks if they wish to.
There is only one stack pointer in a CPU, so how can each thread have its own stack?
A stack can be sitting on disk. A stack can be sitting in memory but not being used as a stack.
And what is the difference b/w stack and stack frame? From what I've read, there is only one stack and frames are pushed on it.
Right, so a single stack could have several frames pushed onto it. When one function finishes, it pops of its stack frame and returns to the caller with the caller's frame on the top of the stack.
Again, it is a physical stack?
I don't know what that means.
Do these stack exist in the virtual memory?
Yes. That's why one thread can easily access variables on another thread's stack if the address is passed from one to the other. A stack is just some memory that's being used as a stack.
Summary of my understanding:
The top memory addresses are used for the? (I initially thought there was only one call stack) stack, and the? stack grows downwards (What and where are the stack and heap?)
However, each thread gets it's own stack allocated, so there should be multiple call stacks in memory (https://stackoverflow.com/a/80113/2415178)
Applications can share threads (e.g, the key application is using the main thread), but several threads can be running at the same time.
There is a CPU register called sp that tracks the stack pointer, the current stack frame of a call stack.
So here's my confusion:
Do all of the call stacks necessary for an application (if this is even possible to know) get allocated when the application gets launched? Or do call stacks get allocated/de-allocated dynamically as applications spin off new threads? And if that is the case, (I know stacks have a fixed size), do the new stacks just get allocated right below the previous stacks-- So you would end up with a stack of stacks in the top addresses of memory? Or am I just fundamentally misunderstanding how call stacks are being created/used?
I am an OS X application developer, so my visual reference for how call stacks are created come from Xcode's stack debugger:
Now I realize that how things are here are more than likely unique to OS X, but I was hoping that conventions would be similar across operating systems.
It appears that each application can execute code on multiple threads, and even spin off new worker threads that belong to the application-- and every thread needs a call stack to keep track of the stack frames.
Which leads me to my last question:
How does the sp register work if there are multiple call stacks? Is it only used for the main call stack? (Presumably the top-most call stack in memory, and associated with the main thread of the OS) [https://stackoverflow.com/a/1213360/2415178]
Do all of the call stacks necessary for an application (if this is even possible to know) get allocated when the application gets launched?
No. Typically, each thread's stack is allocated when that thread is created.
Or do call stacks get allocated/de-allocated dynamically as applications spin off new threads?
Yes.
And if that is the case, (I know stacks have a fixed size), do the new stacks just get allocated right below the previous stacks-- So you would end up with a stack of stacks in the top addresses of memory? Or am I just fundamentally misunderstanding how call stacks are being created/used?
It varies. But the stack just has to be at the top of a large enough chunk of available address space in the memory map for that particular process. It doesn't have to be at the very top. If you need 1MB for the stack, and you have 1MB, you can just reserve that 1MB and have the stack start at the top of it.
How does the sp register work if there are multiple call stacks? Is it only used for the main call stack?
A CPU has as many register sets as threads that can run at a time. When the running thread is switched, the leaving thread's stack pointer is saved and the new thread's stack pointer is restored -- just like all other registers.
There is no "main thread of the OS". There are some kernel threads that do only kernel tasks, but also user-space threads also run in kernel space to run the OS code. Pure kernel threads have their own stacks somewhere in kernel memory. But just like normal threads, it doesn't have to be at the very top, the stack pointer just has to start at the highest address in the chunk used for that stack.
There is no such thing as the "main thread of the OS". Every process has its own set of threads, and those threads are specific to that process, not shared. Typically, at any given point in time, most threads on a system will be suspended awaiting input.
Every thread in a process has its own stack, which is allocated when the thread is created. Most operating systems will leave some space between each stack to allow them to grow if needed, and to prevent them from colliding with each other.
Every thread also has its own set of CPU registers, including a stack pointer (pointing to a location in that thread's stack).
I have two threads created using the CreateThread() function of the Windows API. My first thread (let's call it IMPORTANT_FAST_ACCESS) pushes some data onto a stack. The second thread (let's call it STACK_ACCESS_SLOW) want's to pop some data off that stack and work with it.
The problem is I want the IMPORTANT_FAST_ACCESS thread to be able to access the stack immediately without waiting for the second thread to do it's operation on it. So I basically want the IMPORTANT_FAST_ACCESS to be allowed to access the stack whenever it wants while the STACK_ACCESS_SLOW thread will look for when it is not used by the other thread and grab pop the stack data it wants.
However what if STACK_ACCESS_SLOW starts to access the stack because it was free 1 second ego but now IMPORTANT_FAST_ACCESS decided it wants to push some data on it and cannot wait?
How can I solve this issue?
You have to synchronize access to the shared stack, no matter how fast or slow the threads are. You cannot allow one thread to modify the stack while another thread is using it, and vice versa.
If you don't want the faster thread to be delayed by the slower thread, you have to make sure the slower thread does not keep a lock on the stack while processing it. That means having the slower thread either:
pop stack items one at a time, locking and unlocking the stack each time.
lock the stack, make a copy/swap of it, unlock it, and process the copy as needed.
How is stack space allocated (in the same address space) to each thread of a process in Linux or any other OS for that matter?
It depends on the type of thread library, a user space library like pthreads would allocate memory and divide it into thread stacks. On the OS side each thread would get a kernel stack.
On creation of new thread, the operating system reserves space in stack segment for current thread (parent), where the future auto variables and function call data of parent will live. Then, it allocates one guard page (this is to prevent the parent colliding into child stack, but this may vary with different operating systems). Once this is done, the stack frame for child thread is created (which is typically one-two page(s)).
This process is repeated in case the parent spawns multiple threads. All these stack frames live in stack segment of address space of process whose all these threads are part of.
I just begin to learn OS. I feel puzzle about stack. As I found the stack is attached to each thread. That means the life of the stack is when the thread is created and be reclaimed when the thread is completion.
Also search from the google, the argument and some local variable are stored in the thread. But these are allocated at compile time which seems conflict with the former that the stack is attached to a thread and be reclaimed after the thread is finished.
Any one could give me some detail explanation?
the argument and some local variable are stored in the thread. But these are allocated at compile time
That is not correct.
When a thread is started, a stack is associated with that thread. When a thread terminates, that stack will be reclaimed.
For an example of why that cannot be allocated at compile time, imagine a program that prompts the user for a number of threads to start, and then starts that number of threads. There is no way the compiler could allocate storage for the arguments to methods running on that thread, or for local storage associated with that thread.