In C, linked lists are usually implemented as a series of nodes stored on the heap that point to eachother. The heap is a persistent memory area that runs throughout the life-cycle of the program.
When you create a variable normally in a C function, and the function returns, the variable that you created is no longer accessible. However when you create something on the heap in a function, and the function is returned, the data you allocated on the heap is still there. However, you have no way of accessing it-- unless the function returns a pointer.
So what you would do for create_q() would be to create the linked list on the heap (using a function in stdlib.h called "malloc"), and then you would return a pointer to your first node, letting the main function know where on the heap to find the first node. Then that first node would have a pointer in it, telling the program where on the heap to find the second node, and so forth.
However, you're probably approaching linked lists the wrong way. Unless this is for some sort of homework project, you probably wouldn't want to create an empty linked list. One of the benefits of a linked list is that it's a dynamic structure in which you can easily insert new nodes. You could still have some variable keeping track of the maximum size you want the list to be, but you probably wouldn't want to actually create the nodes until you had to.
Just keep in mind what a linked list is. It's a set of nodes floating on the heap (in C) that each store some data, and contain a pointer to the next node floating on the heap. All you need, to access the linked list, is a pointer to the first node. To add a new node, you simply "walk" through the list till you reach the last node, and then create a new node and have the old-last node point to it.