In this video we'll take a detailed dive into the program's data and the sub segments that make this up. We generally refer to this data memory as the data segment. You spent a great deal of time learning about the different physical memories of a micro controller. Including flash, RAM and Register memories. Each with a different characteristics. These characteristics included things like volatility, speed, access, as well as many others. A microcontroller utilizes these memories in many different ways when a C program is built and installed on an embedded system. A compiled program references two basic blocks of memory, the program code and the program data. These two types are also known as the code and data segments. Both of these segments map compiled segments to a physical address space provided via the linker file. Data memory can be generically described as the operands that are program executable operates on. Data memory can have many different forms for data address information, configuration information, or just raw application data. At a minimum, you should think of the variables that you create in your program as your data memory. This includes derived data types such as pointers, structures, and arrays, in addition to single variables. The contents of your data memory is regularly changing throughout your program's execution. This is done by loading your data from your data memory into CPU registers, operating on it there, and then storing the results back into memory. However, in order to properly describe data memory in the data segment, we will not refer to data as variables but as allocated data. Meaning not all data we create needs to be in the form of a variable. Data can be allocated both at compile time and runtime. In addition data can have different characteristics regardless of your data and the different methods of managing your data we will generally represent it all as being a part of the data segment. The data segment is a method of organising data memory and compile time in order to map address in the physical. The data segment is broken into four main sub segments. These are the stack the heap the data and the BSS regions. Each segment size depends on how you wrote your program. The Stack stores temporary data in most local variables. The Heap stores dynamic information at runtime. The Data region stores non-zero initialized global and static data. And the BSS stores uninitialized and zero initialized global and static data. Let us open up our linker file example. Since our code memory contains our operations that we perform, those operations needs to know exact locations on where the data they are operating on exist. The linker file provides segment and sub-segment mapping to physical memories. Careful management of the code and data at compile time substitute the loading and storing of data into the CPU using addresses and address offsets to unique locations and memory. Here you can see the same sub-segments of the data segment we have spoken about previously. They met to a physical data region of RAM type memory and have a length of 64 kbytes. Without carefully managing these data locations and memory you could easily corrupt your data by mis-addressing a variable or writing outside the bounds of the segment space the linker file will specify in what order they should be placed. Therefore, each sub-segment also has boundaries given either by a link or flag, or by the size of the memory you allocated. It is highly likely that not all the memory will be allocated at compile time. This will just leave some leftover memory unused for your program. If you were to create a physical graph of this memory regions and you can see how they line up contiguously and the address phase. Starting at the origin of addresses and going for the link of the memory. In this example, you can see that the stack was place at a high address. Meaning instead of placing this right after the hit, place it starting at the end of memory region and have a descend back towards the other segments. This type of stack is specified as a full descending stack meaning it starts at a higher address and grows towards lower addresses as more data gets added into it. There are many different characteristics that variables in data can have in your program. Data can have different sizes, allocation times, life time, scope and locations. This method of allocation and management of data are handled in number of different ways, specific CQ words like just Type Qualifiers, Modifiers and Storage Classes can help specify how your variable means to be manage, compiler tributes can also give us extra option to manipulate how data is allocated. There are also some runtime routines that can help us allocate data at runtime. The most obvious characteristics of data is the size, and it's specified by either a C data type or they run time parameter. However, data sizes are architecture and compiler dependent. The C standards specifies minimum size that some types must be. But a compiler in an architecture can manipulate types differently. We will discuss this in more detail later. Lifetime and scope of data are somewhat related, there are three forms of lifetime these include data that exist for the lifetime of a function or a block, data that exist for the lifetime of a program. And data that exist for longer than a function but less than a program duration those local variables only exist for the length of the function or block. This allocated in runtime when the function is called there destroyed when the function returns, this variables have locals scope which means there only accessable during the function of block. The memory they occupy is reused by other functions as a program executes. The only exception to this are local variables created with the static type storage class. Global variables are variables that get allocated at compile time. They have a lifetime of the entire program and are never destroyed. This one time allocation cause global variables to occupy memory that can never be reused during a program. In addition, global variables have a global scope meaning they're accessible by other parts of a program. Other functions are interrupting can openly access a global variable. In addition to global variables memory can be allocated directly in the linker file in referencing your program. And it will be allocated for the lifetime of the program. We will leave this as an advanced programming topic for later in the course. Lifetime and scope vary slightly in the sense that most local variables have a lifetime of a function or a block. And all global variables have a lifetime of the program. There are, however, a third kind of data allocation that can live for longer than a function, but less than a program. This is referred to as Dynamically Allocated Data. This requires special routines to allocate and free data at runtime. There are many issues with managing dynamic memory. However Dynamic Allocation allows you to re-use memory over and over again rather than statically allocating it. It also provides you with the ability to access data outside of a local scope. We will discuss dynamic memory into some more details later. One major issue with data memory is that data stored in RAM is lost when power is removed from the microcontroller. Therefore, all the state and all the processing and execution is lost because RAM is a volatile memory type. Data can also be stored in non-volatile memory such as flash, but it is a slower and more difficult process to write data to flash during the execution of a program. You're likely to see an extra piece of non volatile memory connected to your micro controller for data that needs to be retained between power cycles. Almost all programs have some type of non volatile data they need to store. You have likely written programs with variables that have initial conditions. These are non volatile pieces of data. They need to be initialized at the beginning of execution of a program or a function. A data segment contains many different sub segments that are used for different kinds of data, allocation of data can be at compile time or at run time. Certain parts of the data segment can be reused and others exist for the lifetime of a program. Scope and access to variables can also dictate the section of the data segment where certain pieces of data exist. Next, we'll show how to allocate data in these different segments with your C programs.
