OpenCores
URL https://opencores.org/ocsvn/s6soc/s6soc/trunk

Subversion Repositories s6soc

Compare Revisions

  • This comparison shows the changes necessary to convert path 
    /s6soc/trunk/doc/src
    from Rev 50 to Rev 54
    ↔ Reverse comparison
  •

Rev 50 → Rev 54

/spec.tex
38,11 → 38,12
%%
\usepackage{import}
\usepackage{bytefield}
\usepackage{listings}
\project{CMod S6 SoC}
\title{Specification}
\author{Dan Gisselquist, Ph.D.}
\email{dgisselq (at) opencores.org}
\revision{Rev.~0.3}
\email{dgisselq (at) ieee.org}
\revision{Rev.~0.4}
\begin{document}
\pagestyle{gqtekspecplain}
\titlepage
63,7 → 64,7
with this program. If not, see \texttt{http://www.gnu.org/licenses/} for a copy.
\end{license}
\begin{revisionhistory}
0.5 & 2/13/2017 & Gisselquist & Draft updates to support the 8-bit byte ZipCPU
0.4 & 3/22/2017 & Gisselquist & Updated to support the 8-bit byte ZipCPU
\\\hline
0.3 & 5/23/2016 & Gisselquist & Draft for comment, includes ZipOS and PMod
pin mapping\\\hline
76,16 → 77,16
\listoffigures
\listoftables
\begin{preface}
The Zip CPU was built with the express purpose of being an area optimized,
32--bit FPGA soft processor.
The ZipCPU was built with the express purpose of being an area optimized,
32--bit FPGA soft processor.
 
The S6~SoC is designed to prove that the Zip~CPU has met this goal.
The S6~SoC is designed to prove that the ZipCPU has met this goal.
 
There are two side--effects to this. First, the project proves how capable a
CMod--S6 is, and second, this project provides demonstration implementations
of how to interact with a variety of PMod devices: the audio amplifier,
the serial 2--line LCD display, the USBUART, and the 16--character numeric
keypad.
very small FPGA, such as the Spartan~6/LX4 on the CMod--S6, can be and second,
this project provides demonstration implementations of how to interact with a
variety of PMod devices: the audio amplifier, the serial 2--line LCD display,
the USBUART, and the 16--character numeric keypad.
\end{preface}
 
\chapter{Introduction}
93,19 → 94,19
\setcounter{page}{1}
 
% What is old
The Zip~CPU is a soft core CPU designed to fit within an FPGA, to use a minimum
amount of the FPGA's resources, and yet to provide the services of a fully
capable 32--bit computer. It is based upon a Von~Neumann architecure and so a
single 32--bit wishbone bus provides it with access to both peripherals and
memory.
The ZipCPU is a soft core Central Processing Unit (CPU) designed to fit within
an FPGA, to use a minimum amount of the FPGA's resources, and yet to still
provide the services of a fully capable 32--bit computer. It is based upon a
Von~Neumann architecture and so a single 32--bit wishbone bus provides it with
access to both peripherals and memory.
 
% What does the old lack?
Previous demonstrations of the Zip~CPU have focused on larger FPGAs, such as
Previous demonstrations of the ZipCPU have focused on larger FPGAs, such as
the Spartan--6 LX9 and LX25 and the Artix--7 35T. On these FPGA's,
the Zip~CPU runs in a pipelined configuration where it is tightly integrated
the ZipCPU runs in a pipelined configuration where it is tightly integrated
with a prefetch/instruction cache. While these demonstrations have shown that
the Zip~CPU can act as a very simple CPU in these environments, they really
haven't demonstrated the ability of the Zip~CPU to use only a minimum amount
the ZipCPU can act as a very simple CPU in these environments, they really
haven't demonstrated the ability of the ZipCPU to use only a minimum amount
of the FPGA's resources.
 
% What is new
112,15 → 113,16
The CMod~S6 board provides the opportunity for that demonstration rather nicely.
% What does the new have that the old lacks
\begin{enumerate}
\item The Spartan--6 LX4 FPGA is very limited in its resources:
It only has 2,400 look--up tables (LUTs), and can only support
a 4,096~Word RAM memory (16 kB).
\item With only 4kW RAM, the majority of any program will need to be placed into
and run from flash.
\item The Spartan--6 LX4 FPGA is very limited in its resources:
It only has 2,400 look--up tables (LUTs), and can only support
a 16~kB RAM memory.
\item With only 16kB RAM, the majority of any program will need to be placed
into and run from the flash.
\item While the chip has enough area for the CPU, it doesn't have enough area
to include the CPU and \ldots write access to the flash, debug access,
wishbone command access from the UART, pipelined CPU operations,
a prefetch cache and more. Other solutions will need to be found.
a prefetch cache and more. Other solutions will need to be found as
part of this project.
\end{enumerate}
 
Of course, if someone just wants the functionality of a small, cheap, CPU,
133,25 → 135,28
 
If, on the other hand, what you want is a small, cheap, CPU that can be
embedded within an FPGA without using too much of the FPGA's resources, this
project will demonstrate that possibility as well as some utility.
Alternatively, if you wish to study how to get a CPU to work in a small,
constrained environment, this project may be what you are looking for.
project will demonstrate that possibility as well as showing that the CPU even
has some utility. Alternatively, if you wish to study how to get a CPU to
work in a small, constrained environment, this project may be what you are
looking for. Likewise, if you are a software developer forced to get something
running on an FPGA, such as the CMod-S6, you may have just found your answer.
 
Finally, because the Zip~CPU and the included ZipOS are as small and simple as
Finally, because the ZipCPU and the included ZipOS are as small and simple as
they are, the code base found here will be simpler to understand than the
code bases for some of these other projects. For example, the Zip~CPU
instruction set is very simple. With less than 40 instructions, it is much
code bases for some of these other projects. For example, the ZipCPU
instruction set is very simple. With only 29~instructions, it is much
easier to understand and learn than the ARM instruction set. Further, unlike
the ARM, the entire specification for and description of the Zip~CPU is
the ARM, the entire specification for and description of the ZipCPU is
publicly available. Likewise, an operating system such as the ZipOS that has
less than \hbox{3,000} lines of code will be much easier to understand than any
Linux operating system--even if it has much less functionality.
Linux operating system--even if it has much less functionality.\footnote{I'm
still entertaining thoughts of placing Linux onto this device.}
 
\chapter{Architecture}
Fig.~\ref{fig:architecture}
\begin{figure}\begin{center}
\includegraphics[width=5in]{../gfx/s6bones.eps}
\caption{CMod S6 SoC Architecture: Zip~CPU and
\caption{CMod S6 SoC Architecture: ZipCPU and
Peripherals}\label{fig:architecture}
\end{center}\end{figure}
shows the basic internal architecture of the S6~SoC. In summary, it consists
163,47 → 168,191
external display. All of these devices are then available for the CPU to
interact with.
 
If you are familiar with the Zip CPU, you'll notice this architecture provides
no access to the Zip CPU debug port. There simply wasn't enough room on the
device. Debugging the Zip~CPU will instead need to take place via other means,
such as dumping all registers and/or memory to the serial port on any reboot.
If you are familiar with the ZipCPU, you'll notice this architecture provides
no access to the ZipCPU debug port. There simply wasn't enough room on the
device. Debugging the ZipCPU will instead need to take place via other means,
such as dumping all registers and/or memory to the serial port on any reboot,
and making judicious use of the internal scope.
 
Further, the Zip~CPU has no ability to write to flash memory. For this reason,
there exists an alternate CMod S6~SoC architecture, shown in
Fig.~\ref{fig:altarchitecture}.
Further, the read/write flash controller couldn't fit in the design along
with the ZipCPU, leaving the ZipCPU with a simpler read--only flash controller
and no ability to write to flash memory. For this reason, there exists an
alternate CMod S6~SoC architecture, shown in Fig.~\ref{fig:altarchitecture}.
\begin{figure}\begin{center}
\includegraphics[width=5in]{../gfx/altbones.eps}
\caption{Alternate CMod S6 SoC Architecture: Peripherals, with no
CPU}\label{fig:altarchitecture}
\end{center}\end{figure}
Using this alternate architecture, it should be possible to test the peripherals
and program the flash memory. Both architectures may be loaded into the flash,
together with the programming code for the Zip CPU.
Using this alternate architecture, it is be possible to test the peripherals
and program the flash memory.
 
The basic approach to loading the board is actually quite simple. Using the
Digilent ADEPT JTAG configuration program, {\tt djtgcfg}, the alternate
Digilent ADEPT JTAG configuration program, {\tt djtgcfg}, the alternate
configuration may be written directly to the device. Once this alternate
configuration has been loaded, the flash may be examined and programmed using
the {\tt zipload} utility. This utility uses Digilent's Asynchronous Parallel
Port interface (DEPP) to communicate with the device, and in particular to
Port interface (DEPP) to communicate with the device, and in particular to
tell the device what to write to the flash. When writing to the flash,
the {\tt zipload} utility will program both a primary and an alternate
FPGA configuration into the configuration section of the flash, as well as
computer code into the rest of the flash. Once complete, the system may then be
reloaded with the primary configuration file which will contain an image of
the CPU. The CPU will then begin following the instructions found in flash
memory.
the {\tt zipload} utility can program the FPGA configuration into the
configuration section of the flash, as well as computer code into the rest of
the flash. Once complete, the system may then be reloaded, either by power
down/up or via {\tt djtgcfg}, with the primary configuration file which will
contain an image of the CPU. The CPU will then begin following the
instructions found in flash memory.
 
 
\chapter{Getting Started}
\section{Building the Xilinx BIT file(s)}
The S6SoC nearly fills the CMod-S6 part, so building the BIT files can be a
little tricky, and there are a couple steps to doing this.
 
The first step is to built a Xilinx project, such as in Fig.~\ref{fig:new-prj}.
\begin{figure}
\begin{center}
\includegraphics[width=6in]{../gfx/newprj.eps}
\caption{Create a New Project}\label{fig:new-prj}
\end{center}\end{figure}
So, once you've downloaded the
repository, create a new project. We'll call it ``s6soc'', and declare
the working directory of this project and the xilinx project directory to
be in a {\tt xilinx} subdirectory of the main repository. (You may notice
that bit files already exist in that directory. You can use those, or
continue here.)
 
As this is a new project, you'll need to configure your project as in
Fig.~\ref{fig:projopts}
\begin{figure}
\begin{center}
\includegraphics[width=6in]{../gfx/projopts.eps}
\caption{Create a New Project}\label{fig:projopts}
\end{center}\end{figure}
so as to configure for the proper part. In this case, the proper part comes
from the CMod--S6 schematic, and is shown in the figure.
 
At this point, you'll want to include the Verilog files into the ISE project.
These files can be found in the {\tt rtl/} project subdirectory. While not
all Verilog files are used in all configurations, including unused files is
not a problem. I recommend just including all of them.
 
Once these Verilog files have been included, there are two possible top level
files. The first is the main {\tt toplevel.v} file, used by the SoC itself.
This file includes the ZipCPU as the master of all the peripherals. The
second top level module is found in the {\tt alttop.v} file. This second file
can be used both to test the CMod--S6 hardware as well as to program the flash
with any programs we wish to run in the S6--SoC.
 
Two more files need to be included: {\tt cmod.ucf} and {\tt cmodtop.ucf}. The
first of these, {\tt cmod.ucf}, is the constraint file for the alternate top
file, {\tt alttop.v}. The latter file, {\tt cmodtop.v}, is the constraint file
for the main top level file, {\tt toplevel.v}. You'll want to associate these
two files with their respected top level at this time.
 
Once you have these files included, the next step will be to configure how the
project will be synthesized. To set these options, first select one of the
two top levels to be the top level, and then right click on the
``Synthesize--XST'' to bring up a process properties menu. This menu contains
three sub--screens. If you look at
Figures~\ref{fig:xstopts},
\begin{figure}
\begin{center}
\includegraphics[width=6in]{../gfx/xstopts.eps}
\caption{Synthesis options for low area}\label{fig:xstopts}
\end{center}\end{figure}
%
\ref{fig:hdlopts},
%
\begin{figure}
\begin{center}
\includegraphics[width=6in]{../gfx/hdlopts.eps}
\caption{HDL optimization options for the S6SoC}\label{fig:hdlopts}
\end{center}\end{figure}
%
and~\ref{fig:xpropts},
%
\begin{figure}
\begin{center}
\includegraphics[width=6in]{../gfx/xpropts.eps}
\caption{Xilinx specific options}\label{fig:xpropts}
\end{center}\end{figure}
you can see how I have made these choices to guarantee a minimum LUT count.
 
%
%
%
 
Once you finish configuring the synthesizer, it's time to configure the map
process. This particular menu is within the Implement Design menu, Process
Properties option. Selecting this menu option brings up a dialog, such as
Fig.~\ref{fig:mapopts}.
\begin{figure}
\begin{center}
\includegraphics[width=6in]{../gfx/mapopts.eps}
\caption{Map parameters necessary to optimize the S6SoC}\label{fig:mapopts}
\end{center}\end{figure}
From Fig.~\ref{fig:mapopts}, you can see what optimization settings were
necessary for getting the S6SoC to fit within the CMod--S6 device.
 
Two configuration screens remain. The first, shown in
Fig.~\ref{fig:configopts},
\begin{figure}
\begin{center}
\includegraphics[width=6in]{../gfx/configopts.eps}
\caption{Configuration Options}\label{fig:configopts}
\end{center}\end{figure}
controls how the FPGA will be configured. Important on this screen is the
fact that the CMod--S6 has a Quad SPI flash, and that it can be run from a
clock as fast as 108~MHz. Here, we choose 16~MHz although I know of no reason
why the clock speed could not be made faster.
 
As a final configuration screen, there are options for generating the
programming (bit) file. These are shown in Fig.~\ref{fig:bitfopts}.
\begin{figure}
\begin{center}
\includegraphics[width=6in]{../gfx/bitfopts.eps}
\caption{Programming file (Bit-File) generation options}\label{fig:bitfopts}
\end{center}\end{figure}
From the standpoint of the CMod--S6, the important configuration item here is
that we would like to compress the bit file, so that we can pack as much
information into our flash device as possible.
 
Having gone through all of your configuration options, you may now build the
two configurations, {\tt toplevel.bit} and {\tt alttop.bit} which are used in
this distribution.
 
{\tt make axload} in the main directory (assuming the Digilent ADEPT utilities
are installed) will load the alternate configuration into the flash, while
{\tt make xload} will load the main configuration into the flash. For our
purposes, and until you have a compiled program to run, you will want to load
the alternate configuration into the flash.
 
\section{Building the Compiler}
 
The ZipCPU compiler comes as source from the ZipCPU distribution. To build
this compiler, clone the ZipCPU distribution and follow the build and
installation instructions within it. Ever afterwards, you'll want the ZipCPU
toolchain, {\tt zip-gcc}, {\tt zip-as}, {\tt zip-ld}, etc., in your
path.\footnote{You may need to install several packages to do this, such as:
flex, bison, libbison-dev, libgmp10, libgmp-dev, libmpfr-dev, libmpc-dev,
libelf-dev, ncurses-dev, libelf-dev, and verilator.}
 
%% git clone
%% Setting zip_param_cis to zero
%% Setting ZIP_DIVIDE to zero
%% Setting ZIP_PIPELINED to zero -- Disables the ZIP_ATOMIC instructions
%% Setting ZIP_THUMB to zero -- Disables the CIS instructions
 
 
\section{Building the ADEPT Utilities}
% Install the digilent adept utilities
 
You will also need a copy of Digilent's ADEPT software installed on your
system in order to use the CMod--S6. This includes not only the ADEPT
utilities, but also ADEPT software development kit and it's runtime environment.
We'll need these pieces of software to provide both the {\tt djtgcfg} program
to load the FPGA with our designs, as well as the {\tt DEPP} port, provided by
both the SDK and the runtime environment, necessary to communicate with our
board.
 
% Install the Digilent ADEPT utilities
% Expand the tar file
% tar -xvzf digilent.adept.utilities_2.2.1-x86_64.tar.gz
% Install them
218,90 → 367,133
% Answer Y to install runtime configuration data files into /etc
%
\section{Building and Running the Host Software}
\section{Building the Board Software}
\section{Building the ZipOS}
 
Once you have the ADEPT utilities installed, you may then build the software
in the {\tt sw/host} project subdirectory. The makefile in this directory
has a reference to the location of ADEPT include files, which you may need to
update according to where you have installed ADEPT. You may also need to
adjust the library path, so that it properly references the ADEPT libraries.
(The library path is currently hard coded into the Makefile.)
 
\section{Building the ZipCPU Software}
 
Once you have the ZipCPU toolchain built, you may then build the ZipCPU software
which will be placed onto the board. This software may be found in the
{\tt sw/dev} directory. The ZipOS example can be found in the {\tt sw/zipos}
directory. A {\tt make} command executed in each of these directories should
build the ZipCPU software for the board.
 
\section{Connecting the PMods to your board}
The S6~SoC supports four PMods: The PModUSBUART, the PModAMP2 pwm to audio
amplifier, the PModCLS two line LCD display, and the PModKYPD sixteen
character keypad. These need to be wired up to the CMod~S6 as shown in
Fig.~\ref{fig:pmodplaces} on page~\pageref{fig:pmodplaces}.
 
\section{Loading Software onto your SoC}
Once you have all the software loaded, you may wish to try a simple LED
program. You may find the {\tt sw/dev/blinky.c} program suitable for this
purpose. To run this program, you will need to follow a series of steps:
\begin{enumerate}
\item Build the {\tt sw/dev/blinky} program.
\item Build both Xilinx configurations, {\tt alttop.bit} and {\tt toplevel.bit}.
The main makefile will look for these files in the {\tt xilinx/}
project subdirectory that you made when you created the project.
\item From the main project subdirectory, run {\tt make axload} to load
the alternate configuration onto your board. (You do have your board
plugged in, right?)
\item Then you can run the {\tt zipload} program found in the {\tt sw/host}
directory. {\tt zipload} takes one or two arguments. In this
example, we'll give it a first argument of the {\tt xilinx/toplevel.bit}
file, and a second argument of the program we would like to run,
{\tt sw/dev/blinky}.
\item Assuming {\tt zipload} completed successfully, you can then either run
{\tt make xload} from the main directory, or you may remove and
reconnect power to your board.
\item A blinking light demonstration should now be running on your board.
\end{enumerate}
 
\chapter{Software}
This chapter provides an overview of the software that is available to support
the S6~SoC. This includes not only the RTL, the Makefiles, and the software
that will run on the Zip~CPU within the S6~SoC, but also the support software
that will run on the ZipCPU within the S6~SoC, but also the support software
necessary for communicating with the S6~SoC in its alternate configuration.
 
\section{Directory Structure}
\begin{itemize}
\item[{\tt trunk/bench}] Contains software for emulating the S6 without the S6
present.
\begin{itemize}
\item[{\tt trunk/bench/cpp}] All of the bench testing software is
written in C++, so it is found in this directory. Primary
among these programs is the {\tt zip\_sim} program which will
simulate the Zip~CPU within the S6~SoC. Specifically, it
simulates everything at or below the {\tt busmaster.v} level.
\item[{\tt sim/verilator/}] Contains software for emulating the S6~SoC
without the S6 present.
 
Some, although not all, of the peripherals have been simulated
and made a part of this simulation. These include the
Quad--SPI flash, the UART, and the LED's.
All of the bench testing software is written in C++, so it is found in
this directory. Primary among these programs is the {\tt zip\_sim}
program which will simulate the ZipCPU within the S6~SoC. Specifically,
it simulates everything at or below the {\tt busmaster.v} level.
 
\end{itemize}
\item[{\tt trunk/doc}] All of the documentation for the S6~SoC project may be
Some, although not all, of the peripherals have been simulated and
made a part of this simulation. These include the Quad--SPI flash,
the UART, the LED's and the GPIOs.
 
\item[{\tt doc/}] All of the documentation for the S6~SoC project may be
found in this documentation directory. Specifically, I would commend
your attention to anything with a {\tt .pdf} extension, as these
are the completed documents. Among these you should find a copy of the
GPL copyright under which this software is released, as well as a
GPL copyright under which this software is released, as well as a
pre--built copy of this document.
\begin{itemize}
\item[{\tt trunk/doc/gfx}] Here is where the graphics are located in
\item[{\tt doc/gfx/}] Here is where the graphics are located in
support of this specification document.
\item[{\tt trunk/doc/src}] And here is where the \LaTeX files are
\item[{\tt doc/src/}] And here is where the \LaTeX files are
kept that were used in building both this document as well as
the GPL copyright.
\end{itemize}
\item[{\tt trunk/rtl}] Verilog files. The two top--level files are
{\tt toplevel.v} for the primary top level module, and
{\tt alttop.v} for the alternate load.
\item[{\tt rtl/}] Verilog files. The two top--level files are
{\tt toplevel.v} for the primary top level module, and {\tt alttop.v}
for the alternate load.
\begin{itemize}
\item[{\tt trunk/rtl/cpu}] Verilog files containing the Zip~CPU
core and peripherals. The toplevel file here is the
{\tt zipbones.v} file, although some of the peripherals, such
as the {\tt ziptimer.v} are referenced independently.
\item[{\tt rtl/cpu}] Verilog files containing the ZipCPU core and
peripherals. The toplevel file here is the {\tt zipbones.v}
file, although some of the peripherals, such as the
{\tt ziptimer.v} are referenced independently.
\end{itemize}
\item[{\tt trunk/sw}] The main software directory, primarily a repository
\item[{\tt sw/}] The main software directory, primarily a repository
for software subdirectories beneath it.
\begin{itemize}
\item[{\tt trunk/sw/dev}] This directory holds a variety of
simple programs for the Zip~CPU, such as {\tt helloworld},
{\tt doorbell} and {\tt doorbell2}, as well as software drivers
for various peripherals, such as the real--time clock simulator,
and the keypad and display device drivers.
\item[{\tt trunk/sw/host}] This directory holds support software which
\item[{\tt sw/dev/}] This directory holds a variety of
simple programs for the ZipCPU, such as {\tt blinky},
{\tt helloworld}, {\tt doorbell} and {\tt doorbell2}, as well
as software drivers for various peripherals, such as the
real--time clock simulator, and the keypad and display device
drivers.
 
One key file in this directory is the {\tt cputest.c} file,
which provides a basic test of the CPU and its capabilities.
 
\item[{\tt sw/host/}] This directory holds support software which
can be built on and run on the host machine. Building this
software will involve adjusting the Makefile so that it knows
where your local ADEPT installation directory is.\footnote{Many
of the programs also depend upon the serial number of my CMod
S6 device. This will need to be adjusted in any new install.}
Once built, you will find a variety of very useful programs
within here.
\item[{\tt trunk/sw/zipos}] This directory contains the source code for
a rudimentary, very basic, operating system that I
call the ZipOS.
where your local ADEPT installation directory is. Once built,
you will find a variety of very useful programs within here.
\item[{\tt sw/zipos/}] This directory contains the source code for
a rudimentary, very basic, pre-emptive multitasking operating
system that I call the ZipOS.
\end{itemize}
\end{itemize}
 
\section{Zip CPU Tool Chain}
To build programs for the Zip~CPU, you will need the Zip~CPU toolchain. You
can find this as part of the Zip~CPU project, available at OpenCores. Building
the Zip~CPU project should result in a set of binaries in the
\hbox{\tt zipcpu/trunk/sw/install/cross-tools/bin} directory. Make this
\section{ZipCPU Tool Chain}
To build programs for the ZipCPU, you will need the ZipCPU toolchain. You
can find this as part of the ZipCPU project, available at OpenCores. Building
the ZipCPU project should result in a set of binaries in the
\hbox{\tt sw/install/cross-tools/bin/} subdirectory of your ZipCPU project
directory. Make this
directory a part of your path, and you should be able to build the CMod S6
Zip~CPU software. Specifically, you will need to use {\tt zip-gcc},
ZipCPU software. Specifically, you will need to use {\tt zip-gcc},
{\tt zip-as}, {\tt zip-ld}, and {\tt zip-cpp}. Other tools, such as
{\tt zip-objdump} and {\tt zip-readelf}, may also prove to be very useful when
trying to figure out what is going on within the SoC.
trying to debug what is going on within the SoC.
 
\section{Bench Test Software}
 
Bench testing software currently consists of the {\tt zip\_sim} program found
within {\tt trunk/bench/cpp}. This program requires Verilator to run, and
within {\tt sim/verilator}. This program requires Verilator to run, and
simulates in a cycle accurate fashion, the entire S6~SoC from {\tt busmaster.v}
on down. Further, the external Quad--SPI flash memory, UART, and LED's are
also simulated, although the 2--line display, audio, and keypad are not.
310,23 → 502,20
Several software programs have been built to support the S6~SoC from a nearby
host. These programs include:
\begin{itemize}
\item {\tt dumpuart}: My current approach to debugging involves
dumping the state of the registers and memory to the
UART upon reboot. The dumpuart command found here is
designed to make certain that the UART is first set
up correctly at 9600~Baud, and second that everything
read from the UART is directly sent to both a file and
to the screen. In this fashion, it is similar to the
UNIX {\tt tee} program, save for its serial port
attachment.
\item {\tt readflash}: As I am loathe to remove anything from
a device that came factory installed, the
{\tt readflash} program reads the original installed
configuration from the flash and dumps it to a file.
\item {\tt dumpuart}: My current approach to debugging involves dumping the
state of the registers and memory to the UART upon reboot. The
dumpuart command found here is designed to make certain that the UART
is first set up correctly at 9600~Baud, and second that everything
read from the UART is directly sent to both a file and to the screen.
In this fashion, it is similar to the UNIX {\tt tee} program, save for
its serial port attachment.
\item {\tt readflash}: As I am loathe to remove anything from a device that
came factory installed, the {\tt readflash} program reads the original
installed configuration from the flash and dumps it to a file.
 
This program is only useful when the alternate configuration is loaded.
 
\item {\tt wbregs}: This program offers a capability very similar to the
\item {\tt wbregs}: This program offers a capability very similar to the
PEEK and POKE capability Apple user's may remember from before the
days of Macintosh. {\tt wbregs <address>} will read from the
Wishbone bus the value at the given address. Likewise
339,29 → 528,34
As examples, {\tt wbregs version}, will return the build date, or
version of the RTL. {\tt wbregs spio} reads the special purpose
I/O register, and {\tt wbregs gpio 0xffffffff} will set all output
GPIO ports high while {\tt wbregs gpio 0xffff0000} will set all
GPIO ports high while {\tt wbregs gpio 0xffff0000} will set all
output GPIO ports to ground.
 
This program is only useful when the alternate configuration is loaded.
When the primary {\tt toplevel.v} configuration is loaded, the ZipCPU
will be able to read and write these registers in a similar fashion.
 
\item {\tt zipload}: This is the primary program you will need to get your
software loaded on the CMod. It takes three arguments. The first is
the name of the primary Xilinx configuration file, the second is the
name of an alternate Xilinx configuration file, and the third is the
name of the Zip~CPU program you wish to write to Flash memory.
software loaded on the CMod. It takes two arguments. The first is
the name of the primary Xilinx configuration file, and the second is
the name of the ZipCPU program you wish to write to Flash memory.
 
Each of these arguments is optional. For example, if only one
configuration file is given, the loader will load the primary
configuration. If only one Zip~CPU program is given, the program will
be loaded into the program memory area and the configuration file areas
will be left untouched.
Each of these arguments is optional. For example, if a configuration
file is given, the loader will load the primary configuration. If
a ZipCPU program is given, the program will be loaded into the program
memory area and the configuration file areas will be left untouched.
 
This program is only useful when the alternate configuration is loaded.
As with {\tt wbregs}, this program is only useful when the alternate
configuration is loaded.
\end{itemize}
\section{Zip CPU Programs}
\section{ZipCPU Programs}
The following are a list of simple, independent, single-task example programs
that will run on the S6~SoC:
\begin{itemize}
\item {\tt blinky}: This is a very simple program similar to hello world, with
the difference that the lights have a bit of a different response, the
hello world message is written slower, and interrupts are used to
accomplish this purpose.
\item {\tt helloworld}: The first program any programmer should build,
``Hello, world!'' This program sends the string, ``Hello, world!''
over the UART connection once per second. It is a very valuable
372,6 → 566,21
the primary configuration, and read from the UART port. It also means
that you must have the UART port properly configured and wired to your
CMod board.
 
Unlike other versions of Hello World that you may be familiar with,
this one does not use the C--library. It programs the hardware
directly.
\item {\tt cputest}: This is a simple test of the CPU and all of its
functionality.
The functionality proven is primarily assembly
level, but it does require the C compiler to put its framework
together.
{\tt cputest} is actually one of two programs built from the same
source. The other is named {\tt cputestcis}, and specifically
allows the complex instruction set (CIS) instructions whereas
{\tt cputest} specifically disallows them. The difference, therefore,
is in the command line parameters given to the assembler (via the
compiler).
\item {\tt doorbell}: This annoying program verifies the functionality of the
audio device by playing a doorbell sound to the audio port. It will
then wait ten seconds, and play the doorbell sound again (and again,
378,7 → 587,7
and again). (It gets old after a while ...)
\item {\tt doorbell2}: This adds to the functionality of the {\tt doorbell}
program a wait for keypress, and a display of the current time on the
2--line display. While almost our fully functional program, this
2--line display. While almost our fully functional ZipOS program, this
does not include any menus to configure the device or set time, since
it doesn't include any keypad functionality.
\item {\tt kptest}: A test of whether or not they keypad driver works. When
400,15 → 609,16
rudimentary system calls. In a similar fashion, it does contain just about all
of the functionality necessary for a multi--tasking microcontroller built
around a do--forever loop. For its size, I consider it an impressive
achievement. You are welcome to disagree with me, however.
achievement. You are welcome to disagree with me.
 
This version of the ZipOS starts in the {\tt resetdump.s} code, so that upon
any startup the ZipOS will dump register contents, the BusError register, and
any scope contents to the UART. This can take some time, so you may wish to
configure what you really wish to send--if anything. If desired,
{\tt resetdump} can also be configured to also dump the entire memory as well
while only using 9~memory locations. All of this is quite useful in case the
Zip~CPU encounters a bus error or other sort of error that causes it to hang,
any scope contents to the UART--assuming that the scope had been triggered.
This can take some time, so you may wish to configure what you really wish to
send--if anything. If desired, {\tt resetdump} can be configured to also
dump the entire memory as well while only using ten memory locations in its own
support. All of this is quite useful in case the
ZipCPU encounters a bus error or other sort of error that causes it to hang,
stall, or reboot, as the CPU registers are very carefully not touched prior to
being sent to the UART output port. This extends to all registers save the
supervisor PC and CC registers, which would've been reset by a reboot anyway.
426,7 → 636,7
The user tasks are found (mostly) within {\tt doorbell.c}, also found in the
ZipOS directory. This file contains two kernel entry points, {\tt kntasks()},
which returns the number of user tasks the kernel needs to know about, and
{\tt kinit()}, which builds each of the tasks and connects their file
{\tt kinit()}, which builds each of the tasks and connects their file
descriptors to the various devices they will be referencing.
\subsection{System Calls}
The ZipOS supports a variety of system calls, listed here:
436,7 → 646,7
Halts the execution of a process until an event matching the
{\tt event\_mask} takes place, or a {\tt timeout} (in milliseconds)
has been reached. The events that can take place are a
bitmask of the various interrupts the CPU supports, together with a
bitmask of the various interrupts the CPU supports, together with a
bitmask of the software interrupt values found in {\tt swint.h}.
 
The {\tt timeout} value can either be zero, to return immediately with
452,7 → 662,7
The other thing to be aware of is that events may accumulate before the
wait system call. Nothing within the wait system call clears prior
events. These prior events be returned and cleared, though, if
the wait call indicates an interest in those events.
the wait call indicates an interest in those events.
 
Upon return, the a bitmask of events that have taken place will be
returned to the process.
460,7 → 670,7
\item {\tt int clear(unsigned event\_mask, int timeout)}
 
This system call works closely with the wait system call. Indeed,
when the timeout given is zero, the two function nearly identically.
when the timeout given is zero, the two function nearly identically.
It clears any of the requested events which may be pending, and returns
a bit mask of the events that were pending and cleared.
 
556,68 → 766,56
puts the CPU to sleep while waiting for an interrupt.
 
\chapter{Operation}
The {\tt doorbell} program has been built to illustrate the operation of both
the Zip~CPU, the ZipOS, as well as showing off how all of the various
peripherals work. It was envisioned after my family and I experienced an
unexpected visitor during the wee hours of the morning. The {\tt doorbell}
program is designed to couple the doorbell and the exterior lights to a
single button. Hence, when the doorbell to the house is pressed, the exterior
light is turned on for a half an hour. This, then, would make it difficult
for someone to see inside during this time.
The {\tt doorbell} program found in {\tt sw/zipos|} has been built to
illustrate the operation of both the ZipCPU the ZipOS, as well as showing off
how all of the various peripherals work. It was envisioned after my family
and I experienced an unexpected visitor during the wee hours of the morning.
The {\tt doorbell} program is designed to couple the doorbell and the exterior
lights to a single button. Hence, when the doorbell to the house is pressed,
the exterior light (an LED in the demo) is turned on for a half an hour. This,
then, would make it difficult for someone to see inside during this time.
 
This chapter will present an example of how the {\tt doorbell} program works.
This chapter will present a discussion of how that {\tt doorbell} program works.
 
To run the {\tt doorbell} program, you will first need to build the various
RTL and software support programs just to get the {\tt doorbell} program on
the device:
\begin{enumerate}
\item First build the primary and alternate .bit files by building
with {\tt toplevel.v} and then {\tt alttop.v} as your top--level
RTL files. I like to place my Xilinx work directoy into
{\tt trunk/xilinx}, and if you do the same the load scripts that
are referenced next will work.
\item First build the primary and alternate .bit files by building with
{\tt toplevel.v} and then {\tt alttop.v} as your top--level RTL files.
I like to place my Xilinx work directory into a {\tt xilinx/} project
subdirectory, and if you do the same the load scripts that are
referenced next will work.
 
Before going on, double check that both configuration .bit files were
created, that they each fit within the device (there would be errors
created, and that they each fit within the device (there would be errors
if they did not), and that they met their respective timing
requirements.
 
\item Then, load the alternate bit file into the S6~SoC. You will need
the Digilent tools installed in order to do this. Having done so,
you may run {\tt make axload} from the {\tt trunk/} directory.
If you didn't run the Xilinx ISE from within {\tt tunk/xilinx},
you may run {\tt make axload} from the main project directory.
If you didn't run the Xilinx ISE from within a {\tt xilinx/} project
subdirectory,
you may need to find your .bit files and adjust where they load
from, but this should be fairly straight--forward from the instructions
within the Makefile.
\item Build the software found in the host directory. This sotware depends
\item Build the software found in the host directory. This software depends
upon Digilent's ADEPT toolsuite, so you will need to adjust the
Makefile so that it references the toolsuite.
 
{\em Note:} Currently, the host software is serial--number locked to
my own CMod--S6. Until I fix this and make it more user friendly,
you'll want to switch your software to reference your own CMod.
To do this, look for the {\tt "SN:\ldots"} lines within the `*.cpp'
files. Currently, there is one such line within wbregs, another
within readflash, and the last within zipload. To find our your own
device's serial number, type {\tt djtgcfg enum}. It should find one
(or more) CMod devices connected to your system, and list each of
their serial numbers.
 
Once you've done this, you should have a working set of host support
programs: readflash, wbregs, and zipload.
 
\item Using {\tt wbregs}, you may now test your configuration. {\tt wbregs}
works like the peek and poke programs from a generation ago.
works like the peek and poke programs from a generation ago.
{\tt wbregs <address>} will return the value of the memory (or
peripheral) found at the {\tt <address>}. Some addresses have names,
such as {\tt UART}, {\tt SPIO}, {\tt GPIO}, {\tt PIC}, and so forth.
These names are found in {\tt trunk/sw/host/regdefs.cpp}, and their
mappings in {\tt trunk/sw/host/regdefs.h}.
such as {\tt UART}, {\tt SPIO}, {\tt GPIO}, {\tt PIC}, and so forth.
These names are found in {\tt sw/host/regdefs.cpp}, and their
mappings in {\tt sw/host/regdefs.h}.
 
As examples, if you type {\tt wbregs version} you should be able
to read the version (a.k.a. build date) from the currently installed
.bit file. Likewise if you type {\tt wbregs uart 65}, you should see
an `A' (i.e. a 65) sent from the S6~SoC over the serial port.
an `A' (i.e. a 65) sent from the S6~SoC over the serial port.
{\tt wbregs uart} by itself will read a single character from the
serial port and so on.
 
635,49 → 833,53
You may then wish to change the name of this file, lest you overwrite
it by running {\tt readflash} again later.
 
\item At this point, it's time to build the programs for the Zip~CPU. To do
this, you will first need to download the Zip~CPU project. When
If you wish to restore this bitfile later, load the alternate
configuration and run {\tt zipload qspiflash.bin}.
 
\item At this point, it's time to build the programs for the ZipCPU. To do
this, you will first need to download the ZipCPU project. When
building that project, it will create a directory of programs
(including its compiler) in
{\tt zipcpu/trunk/sw/install/cross-tools/bin}.
(including its compiler) in the {\tt sw/install/cross-tools/bin/}
subdirectory of the ZipCPU project directory.
Include this directory into your path.
 
\item Change into {\tt trunk/sw/dev} to build some device testing files.
{\tt make} by itself should build some of these.
\item Change into the {\tt sw/dev} project subdirectory to build some device
testing files. {\tt make} by itself should build some of these.
 
You should now be ready to run some basic tests on the S6~SoC.
 
\item Let's test the UART first: back out to the main directory {\tt trunk/},
\item Let's test the UART first: back out to the main project directory,
and run\break
{\tt sw/host/zipload sw/dev/helloworld} and then {\tt make xload}.
Now, examine your UART port. (You do have the PModUSBUART installed
and connected, right?) You should see ``Hello, world!'' printed
and connected, right?) You should see ``Hello, World!'' printed
over and over again once each second.
 
\item You may try other test files in a similar manner, such as
{\tt trunk/sw/dev/doorbell} and\break
{\tt trunk/sw/dev/doorbell2}. The first of these will just play the
doorbell over and over again, whereas the second one will wait for a
\item You may try other test files in a similar manner, such as
{\tt sw/dev/doorbell} and\break
{\tt sw/dev/doorbell2}. The first of these will just play the
doorbell over and over again, whereas the second one will wait for a
button press before playing the doorbell sound.
 
\item Now let's go and build the ZipOS together with it's user files. To do
this, enter into the {\tt trunk/sw/zipos} directory and type
this, enter into the {\tt sw/zipos/} directory and type
{\tt make}. If all goes well, you should now have a program named
{\tt trunk/sw/zipos/doorbell} which you can load into your S6~SoC as
{\tt sw/zipos/doorbell} which you can load into your S6~SoC as
well.
 
\item A final load, and we'll be done. To do this, make {\tt axload} again,
and this time {\tt sw/host/zipload xilinx/toplevel.bit sw/zipos/doorbell}.
When you power on your device the next time, or after you
{\tt make xload}, you'll find the ZipOS running on the Zip~CPU.
and this time
{\tt sw/host/zipload xilinx/toplevel.bit sw/zipos/doorbell}.
When you power on your device the next time, or after you
{\tt make xload}, you'll find the ZipOS running on the ZipCPU.
 
\item To test the doorbell, press one of the buttons. You should hear a
doorbell coming out of the PModAMP2 audio port.
 
\item You should also be able to read the time on the LCD display. It will be
the wrong time (the number of seconds since power on \ldots). To set
the correct time, press `A' on the keypad and then type in the 6--digit
time: HHMMSS.
the wrong time (the number of seconds since power on \ldots) initially.
To set the correct time, press `A' on the keypad and then type in the
6--digit time: HHMMSS.
 
If you make a mistake, the `C' key can be used for a backspace.
 
687,7 → 889,7
keypad.
 
\item Now, when the doorbell rings, the LCD will show the time the doorbell
was pressed. If the time is at night, the ourdoor light (oops, I
was pressed. If the time is at night, the outdoor light (oops, I
mean LED\#3) will turn on for a half an hour (currently set to
30~seconds, since I don't have the patience to wait a half hour while
testing).
715,17 → 917,18
while in the alternate configuration.} Furthermore, {\em only} the RAM offers
the capability of byte-wise writes across the bus.
 
This isn't quite true with the other address regions. Accessing the I/O
region, while it will act like a memory, it may also have side-effects. For
example, reading from the debugging scope device's data port will read a word
from the scope's buffer and advance the buffer pointer. (More on that later.)
This isn't quite true with the other address regions. For example, accesses to
the I/O region may also have side-effects--even though this region is accessed
just like memory. For example, reading from the debugging scope device's data
port will read a word from the scope's buffer and advance the buffer pointer.
(More on that later.)
 
Finally, to keep the address decoder simple, many of these addresses are
Finally, to keep the address decoder simple, many of these addresses are
multiply mapped. Hence you may find the I/O peripherals mapped throughout the
{\tt 0x0400}--{\tt 0x07ff} address region. Other memory addresses are similarly
overmapped. This overmapping was a resource minimization feature, to get the
bus to fit within a minimum number of FPGA resources. For this reason,
addresses not explicitly defined in this specification are undefined.
addresses not explicitly defined in this specification are undefined.
Likewise, attempting to write a byte or half-word to anything other than RAM
will have undefined results.
 
733,15 → 936,15
Tbl.~\ref{tbl:ioregs}
\begin{table}[htbp]
\begin{center}\begin{reglist}
PIC &\scalebox{0.8}{\tt 0x0400} & 32 & R/W & Interrupt Controller \\\hline
BUSERR &\scalebox{0.8}{\tt 0x0404} & 32 & R & Last Bus Error Address\\\hline
TIMER &\scalebox{0.8}{\tt 0x0408} & 32 & R/W & ZipTimer\\\hline
WDOG &\scalebox{0.8}{\tt 0x040c} & 32 & R/W & Watchdog Timer\\\hline
PWM &\scalebox{0.8}{\tt 0x0410} & 32 & R/W & PWM Audio Controller\\\hline
SPIO &\scalebox{0.8}{\tt 0x0414} & 32 & R/W & Special Purpose I/O, Keypad, LED Controller \\\hline
GPIO &\scalebox{0.8}{\tt 0x0418} & 32 & R/W & GPIO Controller \\\hline
UART &\scalebox{0.8}{\tt 0x041c} & 32 & R/W & UART data\\\hline
VERSION &\scalebox{0.8}{\tt 0x0420} & 32 & R & Build date\\\hline
PIC &\scalebox{0.8}{\tt 0x0400} & 32 & R/W & Interrupt Controller \\\hline
BUSERR &\scalebox{0.8}{\tt 0x0404} & 32 & R & Last Bus Error Address\\\hline
TIMER &\scalebox{0.8}{\tt 0x0408} & 32 & R/W & ZipTimer\\\hline
WATCHDOG &\scalebox{0.8}{\tt 0x040c} & 32 & R/W & Watchdog Timer\\\hline
PWM &\scalebox{0.8}{\tt 0x0410} & 32 & R/W & PWM Audio Controller\\\hline
SPIO &\scalebox{0.8}{\tt 0x0414} & 32 & R/W & Special Purpose I/O, Keypad, LED Controller \\\hline
GPIO &\scalebox{0.8}{\tt 0x0418} & 32 & R/W & GPIO Controller \\\hline
UART &\scalebox{0.8}{\tt 0x041c} & 32 & R/W & UART data\\\hline
VERSION &\scalebox{0.8}{\tt 0x0420} & 32 & R & Build date\\\hline
\end{reglist}
\caption{I/O Peripheral Registers}\label{tbl:ioregs}
\end{center}\end{table}
762,8 → 965,9
\end{bytefield}
\caption{Programmable Interrupt Control (PIC) Register}\label{fig:picreg}
\end{center}\end{figure}
This controller supports up to fifteen interrupts, however only twelve are
defined within the SoC. These are listed in Tbl.~\ref{tbl:hw-ints}.
This controller supports up to fifteen interrupts, however only eleven are
defined within the SoC, and of those only nine are defined in the primary
configuration. All of these interrupts are listed in Tbl.~\ref{tbl:hw-ints}.
\begin{table}[htbp]
\begin{center}\begin{tabular}{|p{0.9in}|p{0.75in}|p{3.75in}|}\hline
\rowcolor[gray]{0.85} Name & Bit Mask & Description \\\hline\hline
770,10 → 974,11
INT\_BUTTON & 0x001 & A Button has been pressed. \\\hline
INT\_BUSERR & 0x002 & A Wishbone bus error has taken place\\\hline
INT\_SCOPE & 0x004 & The Scope has completed its collection\\\hline
INT\_RTC & 0x008 & An alarm or timer has taken place (assuming the RTC
is installed, and includes both alarm or timer)\\\hline
INT\_TIMA & 0x010 & Timer-A has reached zero\\\hline
INT\_TIMB & 0x020 & Timer-B has reached zero\\\hline
%INT\_RTC & 0x008 & An alarm or timer has taken place (assuming the RTC
% is installed, and includes both alarm or timer)\\\hline
INT\_TIMER & 0x010 & Timer-A has reached zero\\\hline
INT\_TIMB & 0x020 & Timer-B has reached zero. (Alternate config only.)
\\\hline
INT\_UARTRX & 0x040 & A character has been received via the UART\\\hline
INT\_UARTTX & 0x080 & The transmit UART is idle, and ready for its next
character.\\\hline
788,7 → 993,7
\end{center}\end{table}
If any interrupt line is active, the PIC controller
will have that bit set among its active set. Once set, the bit and hence the
interrupt can only be cleared by writing to the controller. Interrupts can
interrupt can only be cleared by writing to the controller. Interrupts can
also be enabled as well. The enabled bit mask controls which interrupt lines
are permitted to interrupt the CPU. Hence, just because an interrupt is active
doesn't mean it will interrupt the CPU--the corresponding bit in the enable
800,7 → 1005,7
To keep operations on this register atomic, most of the bits of this register
have special meanings upon write. The one exception to this is the global
interrupt enable bit. On any write, interrupts will be globally enabled or
disabled based upon the value of this bit. Further, the {\tt ANY} bit is a
disabled based upon the value of this bit. Further, the {\tt ANY} bit is a
read only bit, so writes to it have no effect.
 
Enabling specific interrupts, via writes to the enable lines, are different.
811,8 → 1016,9
To disable a specific interrupt, disable all interrupts and write a one to the
enable line of the interrupt you wish to disable. In this fashion, writing a
{\tt 0x00010000} will disable all interrupts and leave interrupt line zero
disabled when the interrupts are re--enabled later, whereas {\tt 0x07fff0000}
will disable all specific interrupts.
disabled when the interrupts are re--enabled later, whereas {\tt 0x7fff0000}
will not only disable all interrupts, it will leave all all specific interrupts
disabled when interrupts are enabled again later.
 
Interrupts are acknowledged in a fashion similar to enabling interrupts. By
writing a `1' to the active bit mask, the interrupt will be acknowledged and
844,7 → 1050,7
The general purpose ZipTimer can be set to auto reload by setting the top bit
as well as the interval. When so set, the timer will automatically
load it's last set value upon reaching zero and interrupting the CPU. This
effectively turns it into an interrupt timer if desired. To set this feature,
effectively turns it into an interval timer. To set this feature,
write to the timer the number of clock ticks before an interrupt, but also set
the high order bit. In this fashion, writing a {\tt 0x8001387f} will interrupt
the CPU every millisecond, starting one millisecond after the write takes place
860,10 → 1066,10
cause the CPU to reboot if for any reason it locked up and failed to interrupt
at the general timer interrupt request time.
 
The ZipOS uses the ZipTimer for task swapping. By setting the timer for
The ZipOS uses the ZipTimer for task swapping. By setting the timer for
1~ms, the ZipOS examines every task for a potential task swap every millisecond.
Of course, if the various tasks are running from Flash at 52~clocks per
instruction, this means that as few as 1,538~instructions may be executed
Of course, if the various tasks are running from Flash at 17~clocks per
instruction, this means that as few as \hbox{4,705~instructions} may be executed
between timer interrupts, but this can be tuned if necessary for better
performance.
 
884,15 → 1090,18
\end{center}\end{figure}
This controller has been designed for easy writing. To send a sample to the
PWM audio controller, simply write the sample to the controller and clear the
PWM audio interrupt---{\em in that order}. When the audio interrupts the CPU
again, it is ready for the next sample. Do note, however, that the audio
interrupt can only be cleared once a new sample has been written to it.
PWM audio interrupt---{\em in that order}. When the audio controller interrupts
the CPU again, it is ready for the next sample.
Do note, however, that the audio
interrupt can only be cleared once a new sample has been written to it.
Attempts to clear it prior to that will have no effect. (This is why the
order matters.)
 
The audio sample rate has been fixed at 8~kHz. While changing this rate is
easy to do within {\tt busmaster.v}, the rate itself takes some work to keep
up with, so I wouldn't recommend going faster.
The audio sample rate has been fixed at 8~kHz. This rate may be easily
changed within {\tt busmaster.v} by simply setting the number of clocks between
samples (currently set to \hbox{10,000}) and rebuilding the primary
configuration. Whether or not the CPU can keep up with a higher rate might
be application dependent.
 
The audio controller supports two additional functionalities, however. The
first is that the {\tt E} bit will be set upon any read when or if the audio
943,7 → 1152,7
designed to control the on-board LED's, buttons, and keypad. Upon any read,
the register reads the current state of the keypad column output, the keypad
row input, the buttons and the LED's. Writing is more difficult, in order to
make certain that parts of these registers can be modified atomically.
make certain that parts of these registers can be modified atomically.
Specifically, to change an LED, write the new value as well as a `1' to the
corresponding LED change enable bit. The same goes for the keypad column
output, a `1' needs to be written to the corresponding change enable bit in
956,8 → 1165,9
The keypad is a little bit tricker. To wait for a keypad interrupt, one needs
to set the column outputs to zero. To do this, write a {\tt 0x0f00} to the
{\tt SPIO} register. When a user then presses a key, one of the row inputs
will go low and an interrupt will be asserted. The key must then be debounced,
and the ZipOS debounces keys for 5~ms. Once debounced, the key may be read.
will go low and an interrupt will be asserted. The key must then be debounced
in software. The ZipOS accomplishes this by waiting for 5ms.
Once debounced, the key may be read.
To do this, set half of the columns to zero, such as by writing a {\tt 0x0cf00}
to the {\tt SPIO} register. If one of the row values is still zero, then
one of the two columns tested corresponded with the key. This can then be
964,6 → 1174,10
repeated until the correct column has been determined, at which point the
row can be read and the key known.
 
All of this functionality may be checked and tested by loading the alternate
configuration and either reading, {\tt wbregs spio}, or writing
{\tt wbregs spio <value>}.
 
The controller will generate a keypad interrupt whenever any row input is
zero, and a button interrupt whenever any button value is a one. This is a
level triggered interrupt, not edge triggered. What that means is that,
972,13 → 1186,13
to be done in software.
 
\subsection{General Purpose I/O}
The General Purpose Input and Output (GPIO) control register, shown in
The General Purpose Input and Output (GPIO) control register, shown in
Fig.~\ref{fig:gpioreg},
\begin{figure}\begin{center}
\begin{bytefield}[endianness=big]{32}
\bitheader{0-31} \\
\bitbox[lrtb]{16}{Current Input Vals (x16)}\bitbox[lrt]{16}{Current Output} \\
\bitbox[lrtb]{16}{Output Change Enable}\bitbox[lrb]{16}{Values (16-outs)}
\bitbox[lrtb]{16}{Output Change Enable}\bitbox[lrb]{16}{Values (16-outs)}
\end{bytefield}
\caption{GPIO Control Register}\label{fig:gpioreg}
\end{center}\end{figure}
987,9 → 1201,9
the value being placed on the 16--output GPIO pins. To change a GPIO pin,
write the new pin's value to this register, together with setting the
corresponding pin in the bit-mask represented by the upper 16--bits. For
example, to set output pin 0, write a {\tt 0x010001} to the GPIO device. To
clear output pin 0, write a {\tt 0x010000}. Likewise pin two may be set by
writing a {\tt 0x020002}, and both pins may be set by writing {\tt 0x030003},
example, to set output pin zero, write a {\tt 0x010001} to the GPIO device. To
clear output pin zero, write a {\tt 0x010000}. Likewise pin one may be set by
writing a {\tt 0x020002}, and both pins may be set by writing {\tt 0x030003},
etc. This makes it possible to adjust some output pins independent of the
others.
 
1000,7 → 1214,7
controller could be easily adjusted to make such ``do-not-care'' lines if
necessary \ldots).
 
Of the 16 GPIO inputs and the 16 GPIO outputs, two lines have been taken for
Of the 16~GPIO inputs and the 16~GPIO outputs, two lines have been taken for
I2C support, and a third has been stolen to make the PMod's fit on the
board. GPIO line zero, for both input and output, is an I2C data line,
{\tt io\_sda}, and GPIO line one is an I2C clock line, {\tt io\_scl}. If the
1015,16 → 1229,17
Moving on to the UART \ldots although the UART module itself
within the S6~SoC is highly configurable, as built
the UART can only handle 9600~Baud, 8--data bits, no parity, and one stop bit.
Changing this involves changing the constant {\tt UART\_SETUP} within
{\tt toplevel.v}. Further, the UART has only a single byte data buffer, so
Changing this involves changing the constant {\tt UART\_SETUP} within
{\tt toplevel.v}.
Further, the UART has only a single byte data buffer, so
reading from the port has a real--time requirement associated with it: the
data buffer must be emptied before the next value is read.
data buffer must be emptied before the next value is read.
Attempts to read from this port will either return an 8--bit data value from
the port, or if no values are available it will return an {\tt 0x0100}
indicating that fact. In general, reading from the UART port involves first
waiting for the interrupt to be ready, second reading from the port itself,
and then third immediately clearing the interrupt. (The interrupt cannot
be cleared while data is waiting.) Writing to the UART port is done in a
be cleared while data is waiting.) Writing to the UART port is done in a
similar fashion. First, wait until the UART transmit interrupt is asserted
(this will likely be most of the time), second, write to the UART port, and
then third, clear the interrupt. As with the read interrupt, clearing the
1041,13 → 1256,13
 
\section{On-Chip Block RAM}
 
The block RAM is the fastest memory available to the processor. It is also
the {\em only} writeable memory available to the processor. Hence all
non-constant program data {\em must} be placed into block RAM. The Zip~CPU
The block RAM is the fastest memory available to the processor. It is also
the {\em only} writable memory available to the processor. Hence all
non-constant program data {\em must} be placed into block RAM. The ZipCPU
can also run instructions from the block RAM if extra speed is desired. When
runnning from block RAM, the Zip~CPU will nominally take 8~clocks per
instruction, for an effective rate of 10~MIPS. Loads or stores to block RAM
will take one clock longer.
running from block RAM, the ZipCPU will nominally take 6~clocks per
instruction, for an effective rate of 13~MIPS. Loads or stores to block RAM
will take one clock longer.
 
\section{Flash Memory}
The flash memory has been arbitrarily sectioned into three sections, one for
1058,25 → 1273,25
\begin{center}\begin{tabular}{|p{0.75in}|p{0.75in}|p{0.5in}|p{3.0in}|}\hline
\rowcolor[gray]{0.85} Start & End & & Purpose \\\hline\hline
\scalebox{0.9}{\tt 0x1000000} & \scalebox{0.9}{\tt 0x10fffff} & R & Primary configuration space\\\hline
\scalebox{0.9}{\tt 0x1100000} & \scalebox{0.9}{\tt 0x11fffff} & R & Alternate configuration space\\\hline
\scalebox{0.9}{\tt 0x1200000} & \scalebox{0.9}{\tt 0x1ffffff} & R & Zip~CPU program memory\\\hline
% \scalebox{0.9}{\tt 0x1100000} & \scalebox{0.9}{\tt 0x11fffff} & R & Alternate configuration space\\\hline
\scalebox{0.9}{\tt 0x1200000} & \scalebox{0.9}{\tt 0x1ffffff} & R & ZipCPU program memory\\\hline
\end{tabular}
\caption{Flash Address Regions}\label{tbl:flash-addresses}
\end{center}\end{table}
The host program {\tt zipload} can be used to load a Zip~CPU program and
The host program {\tt zipload} can be used to load a ZipCPU program and
configuration files into this address space. To use it, first load the
alternate configuration into the FPGA. Then pass it, as arguments, the
primary, and alternate (if desired), configuration files followed by the
Zip~CPU program file. Then, when the primary configuration is loaded again,
perhaps upon power up, the Zip~CPU will automatically start running from it's
{\tt RESET\_ADDRESS}, {\tt 0x480000}.
ZipCPU program file. Then, when the primary configuration is loaded again,
perhaps upon power up, the ZipCPU will automatically start running from it's
{\tt RESET\_ADDRESS}, {\tt 0x1200000}.
 
When running from Flash memory, the Zip~CPU will nominally take 52~clocks per
instruction, for an effective speed of about 1.5~MIPS.
When running from Flash memory, the ZipCPU will nominally take 17~clocks per
instruction, for an effective speed of about 4.7~MIPS.
 
When using {\tt zipload}, the first bit file argument will load the first
configuration space, the second bit file argument will load the second
configuration space, and the third argument will load the Zip~CPU program
configuration space, and the third argument will load the ZipCPU program
into its space.
 
\chapter{Clocks}
1091,7 → 1306,7
\begin{table}[htbp]
\begin{center}
\begin{portlist}
i\_clk\_8mhz & 1 & Input & Clock\\\hline
i\_clk\_8mhz & 1 & Input & 8~MHz Input Clock\\\hline
o\_qspi\_cs\_n & 1 & Output & Quad SPI Flash chip select\\\hline
o\_qspi\_sck & 1 & Output & Quad SPI Flash clock\\\hline
io\_qspi\_dat & 4 & Input/Output & Four-wire SPI flash data bus\\\hline
1126,7 → 1341,7
could also be controlled via I2C, {\tt io\_sda} and {\tt io\_scl}, although that
is not part of the current demonstration software.
 
The assignment of these pins to external output I/O pins is shown in
The assignment of these pins to external output I/O pins is shown in
Fig.~\ref{fig:physicalio}.
\begin{figure}
\begin{center}

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.