You are here: Home DOCUMENTATION Adapt9S12D Using Your Adapt9S12D Microcontroller Module - Memory Map

Technological Arts Inc.

Your Shopping Cart

Your Cart is currently empty.

Using Your Adapt9S12D Microcontroller Module - Memory Map

Article Index
Using Your Adapt9S12D Microcontroller Module
Overview of features
Getting Started
Setup the hardware
Application Programming
Software Debugging
Software Considerations
Memory Map
Interrupts
Software Considerations - S12 Clock
Hardware Details
Appendix
All Pages

Memory Map

PPAGE and banking are two terms you need to understand relating to addressing memory on the S12.

 Let's start with the CPU (Central Processing Unit).  This is the processor that handles the actual machine code instructions.  It is designed with 16-bit registers, such as X, Y, and PC (Program Counter).  As such, it can address some 65,536 bytes directly (64K), and not one byte more nor less.  It does not care what is on the other end of an address, be it a register, RAM, Flash, or a penguin dancing on a keypad.  An address is just a place to read or write an 8-bit byte.  Every piece of hardware that the CPU communicates with has to map somewhere into this address space.  We will refer to the CPU address space as the Logical Address.  (It makes logical sense to the CPU to find stuff here!)

One consideration to keep in mind regarding memory accesses is that the CPU is designed as a 16-bit processor.  As such, its accesses are optimized when it is doing fetches and writes to even addresses.  (An even address has a zero in the Least Significant Bit position.)  If the CPU has to do a word access with an odd address, then it actually has to make two accesses, and then manipulate the results to produce the expected results.  This increases latency.  So be sure to remember this when setting up tables that will be read frequently.  (Instruction fetches have this issue mostly alleviated by the use of an instruction queue.)

There are the various memory resources that have been included on the S12D chip. The Flash consists of 128K, 256K, or 512K of memory (in DG128, DP256, and DP512 versions, respectively).  Physically, a 512K Flash memory would have its memory cells addressed from $00000 to $3FFFF.  This physical address for Flash is never used by the S12 directly, though.  The only time you really need to be aware of it is if you are preparing an S-record file that will be used by Freescale to create a ROM version of the S12 for OEM use.

Notice that the address space for 512K Flash requires 20 bits to specify, not 16 bits.  Obviously, there is an issue here, as the CPU can't directly access this much memory.  Furthermore, Flash is not the only memory-mapped device we want the CPU to access.  There are also:  RAM, EEPROM, and various I/O (Input/Output) devices, such as that dancing penguin.

What Freescale did to solve this problem was to divide up the CPU logical space into sections, each dedicated to its specific device.  We'll get back to this in a moment.

The next step was to devise a way to determine the value of the extra address lines each piece of hardware had.  To do that there had to be a way of tying all this hardware (such as Flash and RAM), together physically so that the CPU could address them.  The  solution was to define something call the Global address space.  Global addresses start at $000000 and go to $7FFFFF and are the physical address space used to reach real hardware, like Flash or RAM.  It takes 23 bits to define this address space, providing room for 8 Megabytes of stuff.  With this much addressing range, that 512K block of Flash can fit in just fine.  Furthermore, room is reserved for larger Flash devices as well.

The internal devices of the S12 were then placed in the Global Address space at fixed locations.  Notice that Flash is at the top of this address space, and here is where things like PPAGE fit in.  To be able to specify the extra address bits which the CPU can't access, registers were defined to accommodate the values for the additional address lines.  PPAGE is specifically for Flash.  Code can only be run in the Logical Address space of the CPU.  To allow Flash, for example, to be able to run code routines stored in it, it has to be mapped into the Logical Address space of the CPU directly.  This is done by dividing up the Flash into pages of 16K each.  Two of these pages are then fixed into the CPU Logical Address space.  They are always there.  (Okay there are lots of caveats and exceptions I could be going into here, but I'm not going that far for now.)  This way code can be placed in these fixed pages so the CPU has something to tell it what to do when power is first applied.  This is why the interrupt table is in Flash by default.  Space is also allocated in the Logical Address space for a third 16K Flash page.  The hardware Flash page appearing here, though, is determined by the value of the PPAGE register.

For a 512K Flash, the 16K pages are numbered starting at $E0 and going to $FF.  One would think that numbering them should start at $00 and go to $1F, but such is not the case.  This partially has to do with where the Flash is in the Global Address space, namely at the top.  As the possibility exists for larger versions of Flash to become available over time, the space taken by Flash will grow downward.  With 256 pages of 16k bytes each, the maximum possible Flash that can be used by the S12 is 4 megabytes.  If that full amount were available, then the page numbering would go from $00 to $FF.  (With external addressing, a designer could add their own additional memory in this space externally, and then use the PPAGE register to access it.)

The default mapping of Flash to CPU Logical Address space on powerup is as follows:
Flash Bank $FD - CPU Address range is $4000 to $7FFF  (Fixed)
Flash Bank $FE - CPU Address range is $8000 to $BFFF (Window;  Actually Bank determined by PPAGE value.)
Flash Bank $FF - CPU Address range is $C000 to $FFFF (Fixed always)

The PPAGE register is in the CPU Address space at $0030.  Changing the value written here determines the actual 16K Flash bank that is visible to the CPU in the $8000 to $BFFF address range.  This is how the processor can execute code in any Flash bank.  Furthermore, there is an instruction made just for this:  CALL.  With CALL, the programmer can set it up so that the executing code calls a subroutine in another Flash Bank.

Various blocks can be re-mapped via special registers following a reset event, if desired.  The starting address for the RAM block is set via the  INITRM register, the location of the register block is re-mapped via the INITRG register, and the location of the EEPROM block can be re-mapped via the INITEE register..



Last Updated ( Wednesday, 17 April 2019 11:53 )