openocd/doc/openocd.texi

\input texinfo @c -*-texinfo-*-
@c %**start of header
@setfilename openocd.info
@settitle OpenOCD User's Guide
@dircategory Development
@direntry
* OpenOCD: (openocd).      OpenOCD User's Guide
@end direntry
@paragraphindent 0
@c %**end of header

@include version.texi

@copying

This User's Guide documents
release @value{VERSION},
dated @value{UPDATED},
of the Open On-Chip Debugger (OpenOCD).

@itemize @bullet
@item Copyright @copyright{} 2008 The OpenOCD Project
@item Copyright @copyright{} 2007-2008 Spencer Oliver @email{spen@@spen-soft.co.uk}
@item Copyright @copyright{} 2008 Oyvind Harboe @email{oyvind.harboe@@zylin.com}
@item Copyright @copyright{} 2008 Duane Ellis @email{openocd@@duaneellis.com}
@item Copyright @copyright{} 2009 David Brownell
@end itemize

@quotation
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.2 or
any later version published by the Free Software Foundation; with no
Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
Texts.  A copy of the license is included in the section entitled ``GNU
Free Documentation License''.
@end quotation
@end copying

@titlepage
@titlefont{@emph{Open On-Chip Debugger:}}
@sp 1
@title OpenOCD User's Guide
@subtitle for release @value{VERSION}
@subtitle @value{UPDATED}

@page
@vskip 0pt plus 1filll
@insertcopying
@end titlepage

@summarycontents
@contents

@ifnottex
@node Top
@top OpenOCD User's Guide

@insertcopying
@end ifnottex

@menu
* About::                            About OpenOCD
* Developers::                       OpenOCD Developers
* JTAG Hardware Dongles::            JTAG Hardware Dongles
* About JIM-Tcl::                    About JIM-Tcl
* Running::                          Running OpenOCD
* OpenOCD Project Setup::            OpenOCD Project Setup
* Config File Guidelines::           Config File Guidelines
* Daemon Configuration::             Daemon Configuration
* Interface - Dongle Configuration:: Interface - Dongle Configuration
* Reset Configuration::              Reset Configuration
* TAP Declaration::                  TAP Declaration
* CPU Configuration::                CPU Configuration
* Flash Commands::                   Flash Commands
* NAND Flash Commands::              NAND Flash Commands
* PLD/FPGA Commands::                PLD/FPGA Commands
* General Commands::                 General Commands
* Architecture and Core Commands::   Architecture and Core Commands
* JTAG Commands::                    JTAG Commands
* Boundary Scan Commands::           Boundary Scan Commands
* TFTP::                             TFTP
* GDB and OpenOCD::                  Using GDB and OpenOCD
* Tcl Scripting API::                Tcl Scripting API
* FAQ::                              Frequently Asked Questions
* Tcl Crash Course::                 Tcl Crash Course
* License::                          GNU Free Documentation License

@comment DO NOT use the plain word ``Index'', reason: CYGWIN filename
@comment case issue with ``Index.html'' and ``index.html''
@comment Occurs when creating ``--html --no-split'' output
@comment This fix is based on: http://sourceware.org/ml/binutils/2006-05/msg00215.html
* OpenOCD Concept Index::            Concept Index
* Command and Driver Index::         Command and Driver Index
@end menu

@node About
@unnumbered About
@cindex about

OpenOCD was created by Dominic Rath as part of a diploma thesis written at the
University of Applied Sciences Augsburg (@uref{http://www.fh-augsburg.de}).
Since that time, the project has grown into an active open-source project,
supported by a diverse community of software and hardware developers from
around the world.

@section What is OpenOCD?
@cindex TAP
@cindex JTAG

The Open On-Chip Debugger (OpenOCD) aims to provide debugging,
in-system programming and boundary-scan testing for embedded target
devices.

@b{JTAG:} OpenOCD uses a ``hardware interface dongle'' to communicate
with the JTAG (IEEE 1149.1) compliant TAPs on your target board.
A @dfn{TAP} is a ``Test Access Port'', a module which processes
special instructions and data.  TAPs are daisy-chained within and
between chips and boards.

@b{Dongles:} OpenOCD currently supports many types of hardware dongles: USB
based, parallel port based, and other standalone boxes that run
OpenOCD internally. @xref{JTAG Hardware Dongles}.

@b{GDB Debug:} It allows ARM7 (ARM7TDMI and ARM720t), ARM9 (ARM920T,
ARM922T, ARM926EJ--S, ARM966E--S), XScale (PXA25x, IXP42x) and
Cortex-M3 (Stellaris LM3 and ST STM32) based cores to be
debugged via the GDB protocol.

@b{Flash Programing:} Flash writing is supported for external CFI
compatible NOR flashes (Intel and AMD/Spansion command set) and several
internal flashes (LPC1700, LPC2000, AT91SAM7, AT91SAM3U, STR7x, STR9x, LM3, and
STM32x). Preliminary support for various NAND flash controllers
(LPC3180, Orion, S3C24xx, more) controller is included.

@section OpenOCD Web Site

The OpenOCD web site provides the latest public news from the community:

@uref{http://openocd.berlios.de/web/}

@section Latest User's Guide:

The user's guide you are now reading may not be the latest one
available.  A version for more recent code may be available.
Its HTML form is published irregularly at:

@uref{http://openocd.berlios.de/doc/html/index.html}

PDF form is likewise published at:

@uref{http://openocd.berlios.de/doc/pdf/openocd.pdf}

@section OpenOCD User's Forum

There is an OpenOCD forum (phpBB) hosted by SparkFun:

@uref{http://forum.sparkfun.com/viewforum.php?f=18}


@node Developers
@chapter OpenOCD Developer Resources
@cindex developers

If you are interested in improving the state of OpenOCD's debugging and
testing support, new contributions will be welcome.  Motivated developers
can produce new target, flash or interface drivers, improve the
documentation, as well as more conventional bug fixes and enhancements.

The resources in this chapter are available for developers wishing to explore
or expand the OpenOCD source code.

@section OpenOCD GIT Repository

During the 0.3.x release cycle, OpenOCD switched from Subversion to
a GIT repository hosted at SourceForge.  The repository URL is:

@uref{git://openocd.git.sourceforge.net/gitroot/openocd/openocd}

You may prefer to use a mirror and the HTTP protocol:

@uref{http://repo.or.cz/r/openocd.git}

With standard GIT tools, use @command{git clone} to initialize
a local repository, and @command{git pull} to update it.
There are also gitweb pages letting you browse the repository
with a web browser, or download arbitrary snapshots without
needing a GIT client:

@uref{http://openocd.git.sourceforge.net/git/gitweb.cgi?p=openocd/openocd}

@uref{http://repo.or.cz/w/openocd.git}

The @file{README} file contains the instructions for building the project
from the repository or a snapshot.

Developers that want to contribute patches to the OpenOCD system are
@b{strongly} encouraged to work against mainline.
Patches created against older versions may require additional
work from their submitter in order to be updated for newer releases.

@section Doxygen Developer Manual

During the 0.2.x release cycle, the OpenOCD project began
providing a Doxygen reference manual.  This document contains more
technical information about the software internals, development
processes, and similar documentation:

@uref{http://openocd.berlios.de/doc/doxygen/index.html}

This document is a work-in-progress, but contributions would be welcome
to fill in the gaps.  All of the source files are provided in-tree,
listed in the Doxyfile configuration in the top of the source tree.

@section OpenOCD Developer Mailing List

The OpenOCD Developer Mailing List provides the primary means of
communication between developers:

@uref{https://lists.berlios.de/mailman/listinfo/openocd-development}

Discuss and submit patches to this list.
The @file{PATCHES} file contains basic information about how
to prepare patches.


@node JTAG Hardware Dongles
@chapter JTAG Hardware Dongles
@cindex dongles
@cindex FTDI
@cindex wiggler
@cindex zy1000
@cindex printer port
@cindex USB Adapter
@cindex RTCK

Defined: @b{dongle}: A small device that plugins into a computer and serves as
an adapter .... [snip]

In the OpenOCD case, this generally refers to @b{a small adapater} one
attaches to your computer via USB or the Parallel Printer Port.  The
execption being the Zylin ZY1000 which is a small box you attach via
an ethernet cable. The Zylin ZY1000 has the advantage that it does not
require any drivers to be installed on the developer PC. It also has
a built in web interface. It supports RTCK/RCLK or adaptive clocking
and has a built in relay to power cycle targets remotely.


@section Choosing a Dongle

There are several things you should keep in mind when choosing a dongle.

@enumerate
@item @b{Voltage} What voltage is your target - 1.8, 2.8, 3.3, or 5V?
Does your dongle support it?  You might need a level converter.
@item @b{Pinout} What pinout does your target board use?
Does your dongle support it?  You may be able to use jumper
wires, or an "octopus" connector, to convert pinouts.
@item @b{Connection} Does your computer have the USB, printer, or
Ethernet port needed?
@item @b{RTCK} Do you require RTCK? Also known as ``adaptive clocking''
@end enumerate

@section Stand alone Systems

@b{ZY1000} See: @url{http://www.zylin.com/zy1000.html} Technically, not a
dongle, but a standalone box. The ZY1000 has the advantage that it does
not require any drivers installed on the developer PC. It also has
a built in web interface. It supports RTCK/RCLK or adaptive clocking
and has a built in relay to power cycle targets remotely.

@section USB FT2232 Based

There are many USB JTAG dongles on the market, many of them are based
on a chip from ``Future Technology Devices International'' (FTDI)
known as the FTDI FT2232; this is a USB full speed (12 Mbps) chip.
See: @url{http://www.ftdichip.com} for more information.
In summer 2009, USB high speed (480 Mbps) versions of these FTDI
chips are starting to become available in JTAG adapters.

@itemize @bullet
@item @b{usbjtag}
@* Link @url{http://www.hs-augsburg.de/~hhoegl/proj/usbjtag/usbjtag.html}
@item @b{jtagkey}
@* See: @url{http://www.amontec.com/jtagkey.shtml}
@item @b{jtagkey2}
@* See: @url{http://www.amontec.com/jtagkey2.shtml}
@item @b{oocdlink}
@* See: @url{http://www.oocdlink.com} By Joern Kaipf
@item @b{signalyzer}
@* See: @url{http://www.signalyzer.com}
@item @b{evb_lm3s811}
@* See: @url{http://www.luminarymicro.com} - The Stellaris LM3S811 eval board has an FTD2232C chip built in.
@item @b{luminary_icdi}
@* See: @url{http://www.luminarymicro.com} - Luminary In-Circuit Debug Interface (ICDI) Board, included in the Stellaris LM3S9B90 and LM3S9B92 Evaluation Kits.
@item @b{olimex-jtag}
@* See: @url{http://www.olimex.com}
@item @b{flyswatter}
@* See: @url{http://www.tincantools.com}
@item @b{turtelizer2}
@* See:
@uref{http://www.ethernut.de/en/hardware/turtelizer/index.html, Turtelizer 2}, or
@url{http://www.ethernut.de}
@item @b{comstick}
@* Link: @url{http://www.hitex.com/index.php?id=383}
@item @b{stm32stick}
@* Link @url{http://www.hitex.com/stm32-stick}
@item @b{axm0432_jtag}
@* Axiom AXM-0432 Link @url{http://www.axman.com}
@item @b{cortino}
@* Link @url{http://www.hitex.com/index.php?id=cortino}
@end itemize

@section USB JLINK based
There are several OEM versions of the Segger @b{JLINK} adapter. It is
an example of a micro controller based JTAG adapter, it uses an
AT91SAM764 internally.

@itemize @bullet
@item @b{ATMEL SAMICE} Only works with ATMEL chips!
@* Link: @url{http://www.atmel.com/dyn/products/tools_card.asp?tool_id=3892}
@item @b{SEGGER JLINK}
@* Link: @url{http://www.segger.com/jlink.html}
@item @b{IAR J-Link}
@* Link: @url{http://www.iar.com/website1/1.0.1.0/369/1/index.php}
@end itemize

@section USB RLINK based
Raisonance has an adapter called @b{RLink}.  It exists in a stripped-down form on the STM32 Primer, permanently attached to the JTAG lines.  It also exists on the STM32 Primer2, but that is wired for SWD and not JTAG, thus not supported.

@itemize @bullet
@item @b{Raisonance RLink}
@* Link: @url{http://www.raisonance.com/products/RLink.php}
@item @b{STM32 Primer}
@* Link: @url{http://www.stm32circle.com/resources/stm32primer.php}
@item @b{STM32 Primer2}
@* Link: @url{http://www.stm32circle.com/resources/stm32primer2.php}
@end itemize

@section USB Other
@itemize @bullet
@item @b{USBprog}
@* Link: @url{http://www.embedded-projects.net/usbprog} - which uses an Atmel MEGA32 and a UBN9604

@item @b{USB - Presto}
@* Link: @url{http://tools.asix.net/prg_presto.htm}

@item @b{Versaloon-Link}
@* Link: @url{http://www.simonqian.com/en/Versaloon}

@item @b{ARM-JTAG-EW}
@* Link: @url{http://www.olimex.com/dev/arm-jtag-ew.html}
@end itemize

@section IBM PC Parallel Printer Port Based

The two well known ``JTAG Parallel Ports'' cables are the Xilnx DLC5
and the MacGraigor Wiggler. There are many clones and variations of
these on the market.

Note that parallel ports are becoming much less common, so if you
have the choice you should probably avoid these adapters in favor
of USB-based ones.

@itemize @bullet

@item @b{Wiggler} - There are many clones of this.
@* Link: @url{http://www.macraigor.com/wiggler.htm}

@item @b{DLC5} - From XILINX - There are many clones of this
@* Link: Search the web for: ``XILINX DLC5'' - it is no longer
produced, PDF schematics are easily found and it is easy to make.

@item @b{Amontec - JTAG Accelerator}
@* Link: @url{http://www.amontec.com/jtag_accelerator.shtml}

@item @b{GW16402}
@* Link: @url{http://www.gateworks.com/products/avila_accessories/gw16042.php}

@item @b{Wiggler2}
@*@uref{http://www.ccac.rwth-aachen.de/@/~michaels/@/index.php/hardware/@/armjtag,
Improved parallel-port wiggler-style JTAG adapter}

@item @b{Wiggler_ntrst_inverted}
@* Yet another variation - See the source code, src/jtag/parport.c

@item @b{old_amt_wiggler}
@* Unknown - probably not on the market today

@item @b{arm-jtag}
@* Link: Most likely @url{http://www.olimex.com/dev/arm-jtag.html} [another wiggler clone]

@item @b{chameleon}
@* Link: @url{http://www.amontec.com/chameleon.shtml}

@item @b{Triton}
@* Unknown.

@item @b{Lattice}
@* ispDownload from Lattice Semiconductor
@url{http://www.latticesemi.com/lit/docs/@/devtools/dlcable.pdf}

@item @b{flashlink}
@* From ST Microsystems;
@uref{http://www.st.com/stonline/@/products/literature/um/7889.pdf,
FlashLINK JTAG programing cable for PSD and uPSD}

@end itemize

@section Other...
@itemize @bullet

@item @b{ep93xx}
@* An EP93xx based Linux machine using the GPIO pins directly.

@item @b{at91rm9200}
@* Like the EP93xx - but an ATMEL AT91RM9200 based solution using the GPIO pins on the chip.

@end itemize

@node About JIM-Tcl
@chapter About JIM-Tcl
@cindex JIM Tcl
@cindex tcl

OpenOCD includes a small ``Tcl Interpreter'' known as JIM-Tcl.
This programming language provides a simple and extensible
command interpreter.

All commands presented in this Guide are extensions to JIM-Tcl.
You can use them as simple commands, without needing to learn
much of anything about Tcl.
Alternatively, can write Tcl programs with them.

You can learn more about JIM at its website,  @url{http://jim.berlios.de}.

@itemize @bullet
@item @b{JIM vs. Tcl}
@* JIM-TCL is a stripped down version of the well known Tcl language,
which can be found here: @url{http://www.tcl.tk}. JIM-Tcl has far
fewer features. JIM-Tcl is a single .C file and a single .H file and
implements the basic Tcl command set. In contrast: Tcl 8.6 is a
4.2 MB .zip file containing 1540 files.

@item @b{Missing Features}
@* Our practice has been: Add/clone the real Tcl feature if/when
needed. We welcome JIM Tcl improvements, not bloat.

@item @b{Scripts}
@* OpenOCD configuration scripts are JIM Tcl Scripts. OpenOCD's
command interpreter today is a mixture of (newer)
JIM-Tcl commands, and (older) the orginal command interpreter.

@item @b{Commands}
@* At the OpenOCD telnet command line (or via the GDB mon command) one
can type a Tcl for() loop, set variables, etc.
Some of the commands documented in this guide are implemented
as Tcl scripts, from a @file{startup.tcl} file internal to the server.

@item @b{Historical Note}
@* JIM-Tcl was introduced to OpenOCD in spring 2008.

@item @b{Need a crash course in Tcl?}
@*@xref{Tcl Crash Course}.
@end itemize

@node Running
@chapter Running
@cindex command line options
@cindex logfile
@cindex directory search

The @option{--help} option shows:
@verbatim
bash$ openocd --help

--help       | -h       display this help
--version    | -v       display OpenOCD version
--file       | -f       use configuration file <name>
--search     | -s       dir to search for config files and scripts
--debug      | -d       set debug level <0-3>
--log_output | -l       redirect log output to file <name>
--command    | -c       run <command>
--pipe       | -p       use pipes when talking to gdb
@end verbatim

By default OpenOCD reads the configuration file @file{openocd.cfg}.
To specify a different (or multiple)
configuration file, you can use the @option{-f} option. For example:

@example
openocd -f config1.cfg -f config2.cfg -f config3.cfg
@end example

Configuration files and scripts are searched for in
@enumerate
@item the current directory,
@item any search dir specified on the command line using the @option{-s} option,
@item @file{$HOME/.openocd} (not on Windows),
@item the site wide script library @file{$pkgdatadir/site} and
@item the OpenOCD-supplied script library @file{$pkgdatadir/scripts}.
@end enumerate
The first found file with a matching file name will be used.

@section Simple setup, no customization

In the best case, you can use two scripts from one of the script
libraries, hook up your JTAG adapter, and start the server ... and
your JTAG setup will just work "out of the box".  Always try to
start by reusing those scripts, but assume you'll need more
customization even if this works.  @xref{OpenOCD Project Setup}.

If you find a script for your JTAG adapter, and for your board or
target, you may be able to hook up your JTAG adapter then start
the server like:

@example
openocd -f interface/ADAPTER.cfg -f board/MYBOARD.cfg
@end example

You might also need to configure which reset signals are present,
using @option{-c 'reset_config trst_and_srst'} or something similar.
If all goes well you'll see output something like

@example
Open On-Chip Debugger 0.4.0 (2010-01-14-15:06)
For bug reports, read
        http://openocd.berlios.de/doc/doxygen/bugs.html
Info : JTAG tap: lm3s.cpu tap/device found: 0x3ba00477
       (mfg: 0x23b, part: 0xba00, ver: 0x3)
@end example

Seeing that "tap/device found" message, and no warnings, means
the JTAG communication is working.  That's a key milestone, but
you'll probably need more project-specific setup.

@section What OpenOCD does as it starts

OpenOCD starts by processing the configuration commands provided
on the command line or, if there were no @option{-c command} or
@option{-f file.cfg} options given, in @file{openocd.cfg}.
@xref{Configuration Stage}.
At the end of the configuration stage it verifies the JTAG scan
chain defined using those commands; your configuration should
ensure that this always succeeds.
Normally, OpenOCD then starts running as a daemon.
Alternatively, commands may be used to terminate the configuration
stage early, perform work (such as updating some flash memory),
and then shut down without acting as a daemon.

Once OpenOCD starts running as a daemon, it waits for connections from
clients (Telnet, GDB, Other) and processes the commands issued through
those channels.

If you are having problems, you can enable internal debug messages via
the @option{-d} option.

Also it is possible to interleave JIM-Tcl commands w/config scripts using the
@option{-c} command line switch.

To enable debug output (when reporting problems or working on OpenOCD
itself), use the @option{-d} command line switch. This sets the
@option{debug_level} to "3", outputting the most information,
including debug messages. The default setting is "2", outputting only
informational messages, warnings and errors. You can also change this
setting from within a telnet or gdb session using @command{debug_level
<n>} (@pxref{debug_level}).

You can redirect all output from the daemon to a file using the
@option{-l <logfile>} switch.

For details on the @option{-p} option. @xref{Connecting to GDB}.

Note! OpenOCD will launch the GDB & telnet server even if it can not
establish a connection with the target. In general, it is possible for
the JTAG controller to be unresponsive until the target is set up
correctly via e.g. GDB monitor commands in a GDB init script.

@node OpenOCD Project Setup
@chapter OpenOCD Project Setup

To use OpenOCD with your development projects, you need to do more than
just connecting the JTAG adapter hardware (dongle) to your development board
and then starting the OpenOCD server.
You also need to configure that server so that it knows
about that adapter and board, and helps your work.
You may also want to connect OpenOCD to GDB, possibly
using Eclipse or some other GUI.

@section Hooking up the JTAG Adapter

Today's most common case is a dongle with a JTAG cable on one side
(such as a ribbon cable with a 10-pin or 20-pin IDC connector)
and a USB cable on the other.
Instead of USB, some cables use Ethernet;
older ones may use a PC parallel port, or even a serial port.

@enumerate
@item @emph{Start with power to your target board turned off},
and nothing connected to your JTAG adapter.
If you're particularly paranoid, unplug power to the board.
It's important to have the ground signal properly set up,
unless you are using a JTAG adapter which provides
galvanic isolation between the target board and the
debugging host.

@item @emph{Be sure it's the right kind of JTAG connector.}
If your dongle has a 20-pin ARM connector, you need some kind
of adapter (or octopus, see below) to hook it up to
boards using 14-pin or 10-pin connectors ... or to 20-pin
connectors which don't use ARM's pinout.

In the same vein, make sure the voltage levels are compatible.
Not all JTAG adapters have the level shifters needed to work
with 1.2 Volt boards.

@item @emph{Be certain the cable is properly oriented} or you might
damage your board.  In most cases there are only two possible
ways to connect the cable.
Connect the JTAG cable from your adapter to the board.
Be sure it's firmly connected.

In the best case, the connector is keyed to physically
prevent you from inserting it wrong.
This is most often done using a slot on the board's male connector
housing, which must match a key on the JTAG cable's female connector.
If there's no housing, then you must look carefully and
make sure pin 1 on the cable hooks up to pin 1 on the board.
Ribbon cables are frequently all grey except for a wire on one
edge, which is red.  The red wire is pin 1.

Sometimes dongles provide cables where one end is an ``octopus'' of
color coded single-wire connectors, instead of a connector block.
These are great when converting from one JTAG pinout to another,
but are tedious to set up.
Use these with connector pinout diagrams to help you match up the
adapter signals to the right board pins.

@item @emph{Connect the adapter's other end} once the JTAG cable is connected.
A USB, parallel, or serial port connector will go to the host which
you are using to run OpenOCD.
For Ethernet, consult the documentation and your network administrator.

For USB based JTAG adapters you have an easy sanity check at this point:
does the host operating system see the JTAG adapter?  If that host is an
MS-Windows host, you'll need to install a driver before OpenOCD works.

@item @emph{Connect the adapter's power supply, if needed.}
This step is primarily for non-USB adapters,
but sometimes USB adapters need extra power.

@item @emph{Power up the target board.}
Unless you just let the magic smoke escape,
you're now ready to set up the OpenOCD server
so you can use JTAG to work with that board.

@end enumerate

Talk with the OpenOCD server using
telnet (@code{telnet localhost 4444} on many systems) or GDB.
@xref{GDB and OpenOCD}.

@section Project Directory

There are many ways you can configure OpenOCD and start it up.

A simple way to organize them all involves keeping a
single directory for your work with a given board.
When you start OpenOCD from that directory,
it searches there first for configuration files, scripts,
files accessed through semihosting,
and for code you upload to the target board.
It is also the natural place to write files,
such as log files and data you download from the board.

@section Configuration Basics

There are two basic ways of configuring OpenOCD, and
a variety of ways you can mix them.
Think of the difference as just being how you start the server:

@itemize
@item Many @option{-f file} or @option{-c command} options on the command line
@item No options, but a @dfn{user config file}
in the current directory named @file{openocd.cfg}
@end itemize

Here is an example @file{openocd.cfg} file for a setup
using a Signalyzer FT2232-based JTAG adapter to talk to
a board with an Atmel AT91SAM7X256 microcontroller:

@example
source [find interface/signalyzer.cfg]

# GDB can also flash my flash!
gdb_memory_map enable
gdb_flash_program enable

source [find target/sam7x256.cfg]
@end example

Here is the command line equivalent of that configuration:

@example
openocd -f interface/signalyzer.cfg \
        -c "gdb_memory_map enable" \
        -c "gdb_flash_program enable" \
        -f target/sam7x256.cfg
@end example

You could wrap such long command lines in shell scripts,
each supporting a different development task.
One might re-flash the board with a specific firmware version.
Another might set up a particular debugging or run-time environment.

@quotation Important
At this writing (October 2009) the command line method has
problems with how it treats variables.
For example, after @option{-c "set VAR value"}, or doing the
same in a script, the variable @var{VAR} will have no value
that can be tested in a later script.
@end quotation

Here we will focus on the simpler solution:  one user config
file, including basic configuration plus any TCL procedures
to simplify your work.

@section User Config Files
@cindex config file, user
@cindex user config file
@cindex config file, overview

A user configuration file ties together all the parts of a project
in one place.
One of the following will match your situation best:

@itemize
@item Ideally almost everything comes from configuration files
provided by someone else.
For example, OpenOCD distributes a @file{scripts} directory
(probably in @file{/usr/share/openocd/scripts} on Linux).
Board and tool vendors can provide these too, as can individual
user sites; the @option{-s} command line option lets you say
where to find these files.  (@xref{Running}.)
The AT91SAM7X256 example above works this way.

Three main types of non-user configuration file each have their
own subdirectory in the @file{scripts} directory:

@enumerate
@item @b{interface} -- one for each kind of JTAG adapter/dongle
@item @b{board} -- one for each different board
@item @b{target} -- the chips which integrate CPUs and other JTAG TAPs
@end enumerate

Best case:  include just two files, and they handle everything else.
The first is an interface config file.
The second is board-specific, and it sets up the JTAG TAPs and
their GDB targets (by deferring to some @file{target.cfg} file),
declares all flash memory, and leaves you nothing to do except
meet your deadline:

@example
source [find interface/olimex-jtag-tiny.cfg]
source [find board/csb337.cfg]
@end example

Boards with a single microcontroller often won't need more
than the target config file, as in the AT91SAM7X256 example.
That's because there is no external memory (flash, DDR RAM), and
the board differences are encapsulated by application code.

@item Maybe you don't know yet what your board looks like to JTAG.
Once you know the @file{interface.cfg} file to use, you may
need help from OpenOCD to discover what's on the board.
Once you find the TAPs, you can just search for appropriate
configuration files ... or write your own, from the bottom up.
@xref{Autoprobing}.

@item You can often reuse some standard config files but
need to write a few new ones, probably a @file{board.cfg} file.
You will be using commands described later in this User's Guide,
and working with the guidelines in the next chapter.

For example, there may be configuration files for your JTAG adapter
and target chip, but you need a new board-specific config file
giving access to your particular flash chips.
Or you might need to write another target chip configuration file
for a new chip built around the Cortex M3 core.

@quotation Note
When you write new configuration files, please submit
them for inclusion in the next OpenOCD release.
For example, a @file{board/newboard.cfg} file will help the
next users of that board, and a @file{target/newcpu.cfg}
will help support users of any board using that chip.
@end quotation

@item
You may may need to write some C code.
It may be as simple as a supporting a new ft2232 or parport
based dongle; a bit more involved, like a NAND or NOR flash
controller driver; or a big piece of work like supporting
a new chip architecture.
@end itemize

Reuse the existing config files when you can.
Look first in the @file{scripts/boards} area, then @file{scripts/targets}.
You may find a board configuration that's a good example to follow.

When you write config files, separate the reusable parts
(things every user of that interface, chip, or board needs)
from ones specific to your environment and debugging approach.
@itemize

@item
For example, a @code{gdb-attach} event handler that invokes
the @command{reset init} command will interfere with debugging
early boot code, which performs some of the same actions
that the @code{reset-init} event handler does.

@item
Likewise, the @command{arm9 vector_catch} command (or
@cindex vector_catch
its siblings @command{xscale vector_catch}
and @command{cortex_m3 vector_catch}) can be a timesaver
during some debug sessions, but don't make everyone use that either.
Keep those kinds of debugging aids in your user config file,
along with messaging and tracing setup.
(@xref{Software Debug Messages and Tracing}.)

@item
You might need to override some defaults.
For example, you might need to move, shrink, or back up the target's
work area if your application needs much SRAM.

@item
TCP/IP port configuration is another example of something which
is environment-specific, and should only appear in
a user config file.  @xref{TCP/IP Ports}.
@end itemize

@section Project-Specific Utilities

A few project-specific utility
routines may well speed up your work.
Write them, and keep them in your project's user config file.

For example, if you are making a boot loader work on a
board, it's nice to be able to debug the ``after it's
loaded to RAM'' parts separately from the finicky early
code which sets up the DDR RAM controller and clocks.
A script like this one, or a more GDB-aware sibling,
may help:

@example
proc ramboot @{ @} @{
    # Reset, running the target's "reset-init" scripts
    # to initialize clocks and the DDR RAM controller.
    # Leave the CPU halted.
    reset init

    # Load CONFIG_SKIP_LOWLEVEL_INIT version into DDR RAM.
    load_image u-boot.bin 0x20000000

    # Start running.
    resume 0x20000000
@}
@end example

Then once that code is working you will need to make it
boot from NOR flash; a different utility would help.
Alternatively, some developers write to flash using GDB.
(You might use a similar script if you're working with a flash
based microcontroller application instead of a boot loader.)

@example
proc newboot @{ @} @{
    # Reset, leaving the CPU halted.  The "reset-init" event
    # proc gives faster access to the CPU and to NOR flash;
    # "reset halt" would be slower.
    reset init

    # Write standard version of U-Boot into the first two
    # sectors of NOR flash ... the standard version should
    # do the same lowlevel init as "reset-init".
    flash protect 0 0 1 off
    flash erase_sector 0 0 1
    flash write_bank 0 u-boot.bin 0x0
    flash protect 0 0 1 on

    # Reboot from scratch using that new boot loader.
    reset run
@}
@end example

You may need more complicated utility procedures when booting
from NAND.
That often involves an extra bootloader stage,
running from on-chip SRAM to perform DDR RAM setup so it can load
the main bootloader code (which won't fit into that SRAM).

Other helper scripts might be used to write production system images,
involving considerably more than just a three stage bootloader.

@section Target Software Changes

Sometimes you may want to make some small changes to the software
you're developing, to help make JTAG debugging work better.
For example, in C or assembly language code you might
use @code{#ifdef JTAG_DEBUG} (or its converse) around code
handling issues like:

@itemize @bullet

@item @b{ARM Semihosting}...
@cindex ARM semihosting
When linked with a special runtime library provided with many
toolchains@footnote{See chapter 8 "Semihosting" in
@uref{http://infocenter.arm.com/help/topic/com.arm.doc.dui0203i/DUI0203I_rvct_developer_guide.pdf,
ARM DUI 0203I}, the "RealView Compilation Tools Developer Guide".
The CodeSourcery EABI toolchain also includes a semihosting library.},
your target code can use I/O facilities on the debug host.  That library
provides a small set of system calls which are handled by OpenOCD.
It can let the debugger provide your system console and a file system,
helping with early debugging or providing a more capable environment
for sometimes-complex tasks like installing system firmware onto
NAND or SPI flash.

@item @b{ARM Wait-For-Interrupt}...
Many ARM chips synchronize the JTAG clock using the core clock.
Low power states which stop that core clock thus prevent JTAG access.
Idle loops in tasking environments often enter those low power states
via the @code{WFI} instruction (or its coprocessor equivalent, before ARMv7).

You may want to @emph{disable that instruction} in source code,
or otherwise prevent using that state,
to ensure you can get JTAG access at any time.
For example, the OpenOCD @command{halt} command may not
work for an idle processor otherwise.

@item @b{Delay after reset}...
Not all chips have good support for debugger access
right after reset; many LPC2xxx chips have issues here.
Similarly, applications that reconfigure pins used for
JTAG access as they start will also block debugger access.

To work with boards like this, @emph{enable a short delay loop}
the first thing after reset, before "real" startup activities.
For example, one second's delay is usually more than enough
time for a JTAG debugger to attach, so that
early code execution can be debugged
or firmware can be replaced.

@item @b{Debug Communications Channel (DCC)}...
Some processors include mechanisms to send messages over JTAG.
Many ARM cores support these, as do some cores from other vendors.
(OpenOCD may be able to use this DCC internally, speeding up some
operations like writing to memory.)

Your application may want to deliver various debugging messages
over JTAG, by @emph{linking with a small library of code}
provided with OpenOCD and using the utilities there to send
various kinds of message.
@xref{Software Debug Messages and Tracing}.

@end itemize

@node Config File Guidelines
@chapter Config File Guidelines

This chapter is aimed at any user who needs to write a config file,
including developers and integrators of OpenOCD and any user who
needs to get a new board working smoothly.
It provides guidelines for creating those files.

You should find the following directories under @t{$(INSTALLDIR)/scripts},
with files including the ones listed here.
Use them as-is where you can; or as models for new files.
@itemize @bullet
@item @file{interface} ...
think JTAG Dongle. Files that configure JTAG adapters go here.
@example
$ ls interface
arm-jtag-ew.cfg          hitex_str9-comstick.cfg  oocdlink.cfg
arm-usb-ocd.cfg          icebear.cfg              openocd-usb.cfg
at91rm9200.cfg           jlink.cfg                parport.cfg
axm0432.cfg              jtagkey2.cfg             parport_dlc5.cfg
calao-usb-a9260-c01.cfg  jtagkey.cfg              rlink.cfg
calao-usb-a9260-c02.cfg  jtagkey-tiny.cfg         sheevaplug.cfg
calao-usb-a9260.cfg      luminary.cfg             signalyzer.cfg
chameleon.cfg            luminary-icdi.cfg        stm32-stick.cfg
cortino.cfg              luminary-lm3s811.cfg     turtelizer2.cfg
dummy.cfg                olimex-arm-usb-ocd.cfg   usbprog.cfg
flyswatter.cfg           olimex-jtag-tiny.cfg     vsllink.cfg
$
@end example
@item @file{board} ...
think Circuit Board, PWA, PCB, they go by many names.  Board files
contain initialization items that are specific to a board.
They reuse target configuration files, since the same
microprocessor chips are used on many boards,
but support for external parts varies widely.  For
example, the SDRAM initialization sequence for the board, or the type
of external flash and what address it uses.  Any initialization
sequence to enable that external flash or SDRAM should be found in the
board file. Boards may also contain multiple targets:  two CPUs; or
a CPU and an FPGA.
@example
$ ls board
arm_evaluator7t.cfg               keil_mcb1700.cfg
at91rm9200-dk.cfg                 keil_mcb2140.cfg
at91sam9g20-ek.cfg                linksys_nslu2.cfg
atmel_at91sam7s-ek.cfg            logicpd_imx27.cfg
atmel_at91sam9260-ek.cfg          mini2440.cfg
atmel_sam3u_ek.cfg                olimex_LPC2378STK.cfg
crossbow_tech_imote2.cfg          olimex_lpc_h2148.cfg
csb337.cfg                        olimex_sam7_ex256.cfg
csb732.cfg                        olimex_sam9_l9260.cfg
digi_connectcore_wi-9c.cfg        olimex_stm32_h103.cfg
dm355evm.cfg                      omap2420_h4.cfg
dm365evm.cfg                      osk5912.cfg
dm6446evm.cfg                     pic-p32mx.cfg
eir.cfg                           propox_mmnet1001.cfg
ek-lm3s1968.cfg                   pxa255_sst.cfg
ek-lm3s3748.cfg                   sheevaplug.cfg
ek-lm3s811.cfg                    stm3210e_eval.cfg
ek-lm3s9b9x.cfg                   stm32f10x_128k_eval.cfg
hammer.cfg                        str910-eval.cfg
hitex_lpc2929.cfg                 telo.cfg
hitex_stm32-performancestick.cfg  ti_beagleboard.cfg
hitex_str9-comstick.cfg           topas910.cfg
iar_str912_sk.cfg                 topasa900.cfg
imx27ads.cfg                      unknown_at91sam9260.cfg
imx27lnst.cfg                     x300t.cfg
imx31pdk.cfg                      zy1000.cfg
$
@end example
@item @file{target} ...
think chip. The ``target'' directory represents the JTAG TAPs
on a chip
which OpenOCD should control, not a board. Two common types of targets
are ARM chips and FPGA or CPLD chips.
When a chip has multiple TAPs (maybe it has both ARM and DSP cores),
the target config file defines all of them.
@example
$ ls target
aduc702x.cfg     imx27.cfg     pxa255.cfg
ar71xx.cfg       imx31.cfg     pxa270.cfg
at91eb40a.cfg    imx35.cfg     readme.txt
at91r40008.cfg   is5114.cfg    sam7se512.cfg
at91rm9200.cfg   ixp42x.cfg    sam7x256.cfg
at91sam3u1c.cfg  lm3s1968.cfg  samsung_s3c2410.cfg
at91sam3u1e.cfg  lm3s3748.cfg  samsung_s3c2440.cfg
at91sam3u2c.cfg  lm3s6965.cfg  samsung_s3c2450.cfg
at91sam3u2e.cfg  lm3s811.cfg   samsung_s3c4510.cfg
at91sam3u4c.cfg  lm3s9b9x.cfg  samsung_s3c6410.cfg
at91sam3u4e.cfg  lpc1768.cfg   sharp_lh79532.cfg
at91sam3uXX.cfg  lpc2103.cfg   smdk6410.cfg
at91sam7sx.cfg   lpc2124.cfg   smp8634.cfg
at91sam9260.cfg  lpc2129.cfg   stm32.cfg
c100.cfg         lpc2148.cfg   str710.cfg
c100config.tcl   lpc2294.cfg   str730.cfg
c100helper.tcl   lpc2378.cfg   str750.cfg
c100regs.tcl     lpc2478.cfg   str912.cfg
cs351x.cfg       lpc2900.cfg   telo.cfg
davinci.cfg      mega128.cfg   ti_dm355.cfg
dragonite.cfg    netx500.cfg   ti_dm365.cfg
epc9301.cfg      omap2420.cfg  ti_dm6446.cfg
feroceon.cfg     omap3530.cfg  tmpa900.cfg
icepick.cfg      omap5912.cfg  tmpa910.cfg
imx21.cfg        pic32mx.cfg   xba_revA3.cfg
$
@end example
@item @emph{more} ... browse for other library files which may be useful.
For example, there are various generic and CPU-specific utilities.
@end itemize

The @file{openocd.cfg} user config
file may override features in any of the above files by
setting variables before sourcing the target file, or by adding
commands specific to their situation.

@section Interface Config Files

The user config file
should be able to source one of these files with a command like this:

@example
source [find interface/FOOBAR.cfg]
@end example

A preconfigured interface file should exist for every interface in use
today, that said, perhaps some interfaces have only been used by the
sole developer who created it.

A separate chapter gives information about how to set these up.
@xref{Interface - Dongle Configuration}.
Read the OpenOCD source code if you have a new kind of hardware interface
and need to provide a driver for it.

@section Board Config Files
@cindex config file, board
@cindex board config file

The user config file
should be able to source one of these files with a command like this:

@example
source [find board/FOOBAR.cfg]
@end example

The point of a board config file is to package everything
about a given board that user config files need to know.
In summary the board files should contain (if present)

@enumerate
@item One or more @command{source [target/...cfg]} statements
@item NOR flash configuration (@pxref{NOR Configuration})
@item NAND flash configuration (@pxref{NAND Configuration})
@item Target @code{reset} handlers for SDRAM and I/O configuration
@item JTAG adapter reset configuration (@pxref{Reset Configuration})
@item All things that are not ``inside a chip''
@end enumerate

Generic things inside target chips belong in target config files,
not board config files.  So for example a @code{reset-init} event
handler should know board-specific oscillator and PLL parameters,
which it passes to target-specific utility code.

The most complex task of a board config file is creating such a
@code{reset-init} event handler.
Define those handlers last, after you verify the rest of the board
configuration works.

@subsection Communication Between Config files

In addition to target-specific utility code, another way that
board and target config files communicate is by following a
convention on how to use certain variables.

The full Tcl/Tk language supports ``namespaces'', but JIM-Tcl does not.
Thus the rule we follow in OpenOCD is this: Variables that begin with
a leading underscore are temporary in nature, and can be modified and
used at will within a target configuration file.

Complex board config files can do the things like this,
for a board with three chips:

@example
# Chip #1: PXA270 for network side, big endian
set CHIPNAME network
set ENDIAN big
source [find target/pxa270.cfg]
# on return: _TARGETNAME = network.cpu
# other commands can refer to the "network.cpu" target.
$_TARGETNAME configure .... events for this CPU..

# Chip #2: PXA270 for video side, little endian
set CHIPNAME video
set ENDIAN little
source [find target/pxa270.cfg]
# on return: _TARGETNAME = video.cpu
# other commands can refer to the "video.cpu" target.
$_TARGETNAME configure .... events for this CPU..

# Chip #3: Xilinx FPGA for glue logic
set CHIPNAME xilinx
unset ENDIAN
source [find target/spartan3.cfg]
@end example

That example is oversimplified because it doesn't show any flash memory,
or the @code{reset-init} event handlers to initialize external DRAM
or (assuming it needs it) load a configuration into the FPGA.
Such features are usually needed for low-level work with many boards,
where ``low level'' implies that the board initialization software may
not be working.  (That's a common reason to need JTAG tools.  Another
is to enable working with microcontroller-based systems, which often
have no debugging support except a JTAG connector.)

Target config files may also export utility functions to board and user
config files.  Such functions should use name prefixes, to help avoid
naming collisions.

Board files could also accept input variables from user config files.
For example, there might be a @code{J4_JUMPER} setting used to identify
what kind of flash memory a development board is using, or how to set
up other clocks and peripherals.

@subsection Variable Naming Convention
@cindex variable names

Most boards have only one instance of a chip.
However, it should be easy to create a board with more than
one such chip (as shown above).
Accordingly, we encourage these conventions for naming
variables associated with different @file{target.cfg} files,
to promote consistency and
so that board files can override target defaults.

Inputs to target config files include:

@itemize @bullet
@item @code{CHIPNAME} ...
This gives a name to the overall chip, and is used as part of
tap identifier dotted names.
While the default is normally provided by the chip manufacturer,
board files may need to distinguish between instances of a chip.
@item @code{ENDIAN} ...
By default @option{little} - although chips may hard-wire @option{big}.
Chips that can't change endianness don't need to use this variable.
@item @code{CPUTAPID} ...
When OpenOCD examines the JTAG chain, it can be told verify the
chips against the JTAG IDCODE register.
The target file will hold one or more defaults, but sometimes the
chip in a board will use a different ID (perhaps a newer revision).
@end itemize

Outputs from target config files include:

@itemize @bullet
@item @code{_TARGETNAME} ...
By convention, this variable is created by the target configuration
script. The board configuration file may make use of this variable to
configure things like a ``reset init'' script, or other things
specific to that board and that target.
If the chip has 2 targets, the names are @code{_TARGETNAME0},
@code{_TARGETNAME1}, ... etc.
@end itemize

@subsection The reset-init Event Handler
@cindex event, reset-init
@cindex reset-init handler

Board config files run in the OpenOCD configuration stage;
they can't use TAPs or targets, since they haven't been
fully set up yet.
This means you can't write memory or access chip registers;
you can't even verify that a flash chip is present.
That's done later in event handlers, of which the target @code{reset-init}
handler is one of the most important.

Except on microcontrollers, the basic job of @code{reset-init} event
handlers is setting up flash and DRAM, as normally handled by boot loaders.
Microcontrollers rarely use boot loaders; they run right out of their
on-chip flash and SRAM memory.  But they may want to use one of these
handlers too, if just for developer convenience.

@quotation Note
Because this is so very board-specific, and chip-specific, no examples
are included here.
Instead, look at the board config files distributed with OpenOCD.
If you have a boot loader, its source code will help; so will
configuration files for other JTAG tools
(@pxref{Translating Configuration Files}).
@end quotation

Some of this code could probably be shared between different boards.
For example, setting up a DRAM controller often doesn't differ by
much except the bus width (16 bits or 32?) and memory timings, so a
reusable TCL procedure loaded by the @file{target.cfg} file might take
those as parameters.
Similarly with oscillator, PLL, and clock setup;
and disabling the watchdog.
Structure the code cleanly, and provide comments to help
the next developer doing such work.
(@emph{You might be that next person} trying to reuse init code!)

The last thing normally done in a @code{reset-init} handler is probing
whatever flash memory was configured.  For most chips that needs to be
done while the associated target is halted, either because JTAG memory
access uses the CPU or to prevent conflicting CPU access.

@subsection JTAG Clock Rate

Before your @code{reset-init} handler has set up
the PLLs and clocking, you may need to run with
a low JTAG clock rate.
@xref{JTAG Speed}.
Then you'd increase that rate after your handler has
made it possible to use the faster JTAG clock.
When the initial low speed is board-specific, for example
because it depends on a board-specific oscillator speed, then
you should probably set it up in the board config file;
if it's target-specific, it belongs in the target config file.

For most ARM-based processors the fastest JTAG clock@footnote{A FAQ
@uref{http://www.arm.com/support/faqdev/4170.html} gives details.}
is one sixth of the CPU clock; or one eighth for ARM11 cores.
Consult chip documentation to determine the peak JTAG clock rate,
which might be less than that.

@quotation Warning
On most ARMs, JTAG clock detection is coupled to the core clock, so
software using a @option{wait for interrupt} operation blocks JTAG access.
Adaptive clocking provides a partial workaround, but a more complete
solution just avoids using that instruction with JTAG debuggers.
@end quotation

If the board supports adaptive clocking, use the @command{jtag_rclk}
command, in case your board is used with JTAG adapter which
also supports it.  Otherwise use @command{jtag_khz}.
Set the slow rate at the beginning of the reset sequence,
and the faster rate as soon as the clocks are at full speed.

@section Target Config Files
@cindex config file, target
@cindex target config file

Board config files communicate with target config files using
naming conventions as described above, and may source one or
more target config files like this:

@example
source [find target/FOOBAR.cfg]
@end example

The point of a target config file is to package everything
about a given chip that board config files need to know.
In summary the target files should contain

@enumerate
@item Set defaults
@item Add TAPs to the scan chain
@item Add CPU targets (includes GDB support)
@item CPU/Chip/CPU-Core specific features
@item On-Chip flash
@end enumerate

As a rule of thumb, a target file sets up only one chip.
For a microcontroller, that will often include a single TAP,
which is a CPU needing a GDB target, and its on-chip flash.

More complex chips may include multiple TAPs, and the target
config file may need to define them all before OpenOCD
can talk to the chip.
For example, some phone chips have JTAG scan chains that include
an ARM core for operating system use, a DSP,
another ARM core embedded in an image processing engine,
and other processing engines.

@subsection Default Value Boiler Plate Code

All target configuration files should start with code like this,
letting board config files express environment-specific
differences in how things should be set up.

@example
# Boards may override chip names, perhaps based on role,
# but the default should match what the vendor uses
if @{ [info exists CHIPNAME] @} @{
   set  _CHIPNAME $CHIPNAME
@} else @{
   set  _CHIPNAME sam7x256
@}

# ONLY use ENDIAN with targets that can change it.
if @{ [info exists ENDIAN] @} @{
   set  _ENDIAN $ENDIAN
@} else @{
   set  _ENDIAN little
@}

# TAP identifiers may change as chips mature, for example with
# new revision fields (the "3" here).  Pick a good default; you
# can pass several such identifiers to the "jtag newtap" command.
if @{ [info exists CPUTAPID ] @} @{
   set _CPUTAPID $CPUTAPID
@} else @{
   set _CPUTAPID 0x3f0f0f0f
@}
@end example
@c but 0x3f0f0f0f is for an str73x part ...

@emph{Remember:} Board config files may include multiple target
config files, or the same target file multiple times
(changing at least @code{CHIPNAME}).

Likewise, the target configuration file should define
@code{_TARGETNAME} (or @code{_TARGETNAME0} etc) and
use it later on when defining debug targets:

@example
set _TARGETNAME $_CHIPNAME.cpu
target create $_TARGETNAME arm7tdmi -chain-position $_TARGETNAME
@end example

@subsection Adding TAPs to the Scan Chain
After the ``defaults'' are set up,
add the TAPs on each chip to the JTAG scan chain.
@xref{TAP Declaration}, and the naming convention
for taps.

In the simplest case the chip has only one TAP,
probably for a CPU or FPGA.
The config file for the Atmel AT91SAM7X256
looks (in part) like this:

@example
jtag newtap $_CHIPNAME cpu -irlen 4 -expected-id $_CPUTAPID
@end example

A board with two such at91sam7 chips would be able
to source such a config file twice, with different
values for @code{CHIPNAME}, so
it adds a different TAP each time.

If there are nonzero @option{-expected-id} values,
OpenOCD attempts to verify the actual tap id against those values.
It will issue error messages if there is mismatch, which
can help to pinpoint problems in OpenOCD configurations.

@example
JTAG tap: sam7x256.cpu tap/device found: 0x3f0f0f0f
                (Manufacturer: 0x787, Part: 0xf0f0, Version: 0x3)
ERROR: Tap: sam7x256.cpu - Expected id: 0x12345678, Got: 0x3f0f0f0f
ERROR: expected: mfg: 0x33c, part: 0x2345, ver: 0x1
ERROR:      got: mfg: 0x787, part: 0xf0f0, ver: 0x3
@end example

There are more complex examples too, with chips that have
multiple TAPs.  Ones worth looking at include:

@itemize
@item @file{target/omap3530.cfg} -- with disabled ARM and DSP,
plus a JRC to enable them
@item @file{target/str912.cfg} -- with flash, CPU, and boundary scan
@item @file{target/ti_dm355.cfg} -- with ETM, ARM, and JRC (this JRC
is not currently used)
@end itemize

@subsection Add CPU targets

After adding a TAP for a CPU, you should set it up so that
GDB and other commands can use it.
@xref{CPU Configuration}.
For the at91sam7 example above, the command can look like this;
note that @code{$_ENDIAN} is not needed, since OpenOCD defaults
to little endian, and this chip doesn't support changing that.

@example
set _TARGETNAME $_CHIPNAME.cpu
target create $_TARGETNAME arm7tdmi -chain-position $_TARGETNAME
@end example

Work areas are small RAM areas associated with CPU targets.
They are used by OpenOCD to speed up downloads,
and to download small snippets of code to program flash chips.
If the chip includes a form of ``on-chip-ram'' - and many do - define
a work area if you can.
Again using the at91sam7 as an example, this can look like:

@example
$_TARGETNAME configure -work-area-phys 0x00200000 \
             -work-area-size 0x4000 -work-area-backup 0
@end example

@subsection Chip Reset Setup

As a rule, you should put the @command{reset_config} command
into the board file.  Most things you think you know about a
chip can be tweaked by the board.

Some chips have specific ways the TRST and SRST signals are
managed. In the unusual case that these are @emph{chip specific}
and can never be changed by board wiring, they could go here.
For example, some chips can't support JTAG debugging without
both signals.

Provide a @code{reset-assert} event handler if you can.
Such a handler uses JTAG operations to reset the target,
letting this target config be used in systems which don't
provide the optional SRST signal, or on systems where you
don't want to reset all targets at once.
Such a handler might write to chip registers to force a reset,
use a JRC to do that (preferable -- the target may be wedged!),
or force a watchdog timer to trigger.
(For Cortex-M3 targets, this is not necessary.  The target
driver knows how to use trigger an NVIC reset when SRST is
not available.)

Some chips need special attention during reset handling if
they're going to be used with JTAG.
An example might be needing to send some commands right
after the target's TAP has been reset, providing a
@code{reset-deassert-post} event handler that writes a chip
register to report that JTAG debugging is being done.
Another would be reconfiguring the watchdog so that it stops
counting while the core is halted in the debugger.

JTAG clocking constraints often change during reset, and in
some cases target config files (rather than board config files)
are the right places to handle some of those issues.
For example, immediately after reset most chips run using a
slower clock than they will use later.
That means that after reset (and potentially, as OpenOCD
first starts up) they must use a slower JTAG clock rate
than they will use later.
@xref{JTAG Speed}.

@quotation Important
When you are debugging code that runs right after chip
reset, getting these issues right is critical.
In particular, if you see intermittent failures when
OpenOCD verifies the scan chain after reset,
look at how you are setting up JTAG clocking.
@end quotation

@subsection ARM Core Specific Hacks

If the chip has a DCC, enable it. If the chip is an ARM9 with some
special high speed download features - enable it.

If present, the MMU, the MPU and the CACHE should be disabled.

Some ARM cores are equipped with trace support, which permits
examination of the instruction and data bus activity.  Trace
activity is controlled through an ``Embedded Trace Module'' (ETM)
on one of the core's scan chains.  The ETM emits voluminous data
through a ``trace port''.  (@xref{ARM Hardware Tracing}.)
If you are using an external trace port,
configure it in your board config file.
If you are using an on-chip ``Embedded Trace Buffer'' (ETB),
configure it in your target config file.

@example
etm config $_TARGETNAME 16 normal full etb
etb config $_TARGETNAME $_CHIPNAME.etb
@end example

@subsection Internal Flash Configuration

This applies @b{ONLY TO MICROCONTROLLERS} that have flash built in.

@b{Never ever} in the ``target configuration file'' define any type of
flash that is external to the chip. (For example a BOOT flash on
Chip Select 0.) Such flash information goes in a board file - not
the TARGET (chip) file.

Examples:
@itemize @bullet
@item at91sam7x256 - has 256K flash YES enable it.
@item str912 - has flash internal YES enable it.
@item imx27 - uses boot flash on CS0 - it goes in the board file.
@item pxa270 - again - CS0 flash - it goes in the board file.
@end itemize

@anchor{Translating Configuration Files}
@section Translating Configuration Files
@cindex translation
If you have a configuration file for another hardware debugger
or toolset (Abatron, BDI2000, BDI3000, CCS,
Lauterbach, Segger, Macraigor, etc.), translating
it into OpenOCD syntax is often quite straightforward. The most tricky
part of creating a configuration script is oftentimes the reset init
sequence where e.g. PLLs, DRAM and the like is set up.

One trick that you can use when translating is to write small
Tcl procedures to translate the syntax into OpenOCD syntax. This
can avoid manual translation errors and make it easier to
convert other scripts later on.

Example of transforming quirky arguments to a simple search and
replace job:

@example
#   Lauterbach syntax(?)
#
#       Data.Set c15:0x042f %long 0x40000015
#
#   OpenOCD syntax when using procedure below.
#
#       setc15 0x01 0x00050078

proc setc15 @{regs value@} @{
    global TARGETNAME

    echo [format "set p15 0x%04x, 0x%08x" $regs $value]

    arm mcr 15 [expr ($regs>>12)&0x7] \
        [expr ($regs>>0)&0xf] [expr ($regs>>4)&0xf] \
        [expr ($regs>>8)&0x7] $value
@}
@end example



@node Daemon Configuration
@chapter Daemon Configuration
@cindex initialization
The commands here are commonly found in the openocd.cfg file and are
used to specify what TCP/IP ports are used, and how GDB should be
supported.

@anchor{Configuration Stage}
@section Configuration Stage
@cindex configuration stage
@cindex config command

When the OpenOCD server process starts up, it enters a
@emph{configuration stage} which is the only time that
certain commands, @emph{configuration commands}, may be issued.
In this manual, the definition of a configuration command is
presented as a @emph{Config Command}, not as a @emph{Command}
which may be issued interactively.

Those configuration commands include declaration of TAPs,
flash banks,
the interface used for JTAG communication,
and other basic setup.
The server must leave the configuration stage before it
may access or activate TAPs.
After it leaves this stage, configuration commands may no
longer be issued.

@section Entering the Run Stage

The first thing OpenOCD does after leaving the configuration
stage is to verify that it can talk to the scan chain
(list of TAPs) which has been configured.
It will warn if it doesn't find TAPs it expects to find,
or finds TAPs that aren't supposed to be there.
You should see no errors at this point.
If you see errors, resolve them by correcting the
commands you used to configure the server.
Common errors include using an initial JTAG speed that's too
fast, and not providing the right IDCODE values for the TAPs
on the scan chain.

Once OpenOCD has entered the run stage, a number of commands
become available.
A number of these relate to the debug targets you may have declared.
For example, the @command{mww} command will not be available until
a target has been successfuly instantiated.
If you want to use those commands, you may need to force
entry to the run stage.

@deffn {Config Command} init
This command terminates the configuration stage and
enters the run stage.  This helps when you need to have
the startup scripts manage tasks such as resetting the target,
programming flash, etc. To reset the CPU upon startup, add "init" and
"reset" at the end of the config script or at the end of the OpenOCD
command line using the @option{-c} command line switch.

If this command does not appear in any startup/configuration file
OpenOCD executes the command for you after processing all
configuration files and/or command line options.

@b{NOTE:} This command normally occurs at or near the end of your
openocd.cfg file to force OpenOCD to ``initialize'' and make the
targets ready. For example: If your openocd.cfg file needs to
read/write memory on your target, @command{init} must occur before
the memory read/write commands.  This includes @command{nand probe}.
@end deffn

@deffn {Overridable Procedure} jtag_init
This is invoked at server startup to verify that it can talk
to the scan chain (list of TAPs) which has been configured.

The default implementation first tries @command{jtag arp_init},
which uses only a lightweight JTAG reset before examining the
scan chain.
If that fails, it tries again, using a harder reset
from the overridable procedure @command{init_reset}.

Implementations must have verified the JTAG scan chain before
they return.
This is done by calling @command{jtag arp_init}
(or @command{jtag arp_init-reset}).
@end deffn

@anchor{TCP/IP Ports}
@section TCP/IP Ports
@cindex TCP port
@cindex server
@cindex port
@cindex security
The OpenOCD server accepts remote commands in several syntaxes.
Each syntax uses a different TCP/IP port, which you may specify
only during configuration (before those ports are opened).

For reasons including security, you may wish to prevent remote
access using one or more of these ports.
In such cases, just specify the relevant port number as zero.
If you disable all access through TCP/IP, you will need to
use the command line @option{-pipe} option.

@deffn {Command} gdb_port (number)
@cindex GDB server
Specify or query the first port used for incoming GDB connections.
The GDB port for the
first target will be gdb_port, the second target will listen on gdb_port + 1, and so on.
When not specified during the configuration stage,
the port @var{number} defaults to 3333.
When specified as zero, this port is not activated.
@end deffn

@deffn {Command} tcl_port (number)
Specify or query the port used for a simplified RPC
connection that can be used by clients to issue TCL commands and get the
output from the Tcl engine.
Intended as a machine interface.
When not specified during the configuration stage,
the port @var{number} defaults to 6666.
When specified as zero, this port is not activated.
@end deffn

@deffn {Command} telnet_port (number)
Specify or query the
port on which to listen for incoming telnet connections.
This port is intended for interaction with one human through TCL commands.
When not specified during the configuration stage,
the port @var{number} defaults to 4444.
When specified as zero, this port is not activated.
@end deffn

@anchor{GDB Configuration}
@section GDB Configuration
@cindex GDB
@cindex GDB configuration
You can reconfigure some GDB behaviors if needed.
The ones listed here are static and global.
@xref{Target Configuration}, about configuring individual targets.
@xref{Target Events}, about configuring target-specific event handling.

@anchor{gdb_breakpoint_override}
@deffn {Command} gdb_breakpoint_override [@option{hard}|@option{soft}|@option{disable}]
Force breakpoint type for gdb @command{break} commands.
This option supports GDB GUIs which don't
distinguish hard versus soft breakpoints, if the default OpenOCD and
GDB behaviour is not sufficient.  GDB normally uses hardware
breakpoints if the memory map has been set up for flash regions.
@end deffn

@anchor{gdb_flash_program}
@deffn {Config Command} gdb_flash_program (@option{enable}|@option{disable})
Set to @option{enable} to cause OpenOCD to program the flash memory when a
vFlash packet is received.
The default behaviour is @option{enable}.
@end deffn

@deffn {Config Command} gdb_memory_map (@option{enable}|@option{disable})
Set to @option{enable} to cause OpenOCD to send the memory configuration to GDB when
requested. GDB will then know when to set hardware breakpoints, and program flash
using the GDB load command. @command{gdb_flash_program enable} must also be enabled
for flash programming to work.
Default behaviour is @option{enable}.
@xref{gdb_flash_program}.
@end deffn

@deffn {Config Command} gdb_report_data_abort (@option{enable}|@option{disable})
Specifies whether data aborts cause an error to be reported
by GDB memory read packets.
The default behaviour is @option{disable};
use @option{enable} see these errors reported.
@end deffn

@anchor{Event Polling}
@section Event Polling

Hardware debuggers are parts of asynchronous systems,
where significant events can happen at any time.
The OpenOCD server needs to detect some of these events,
so it can report them to through TCL command line
or to GDB.

Examples of such events include:

@itemize
@item One of the targets can stop running ... maybe it triggers
a code breakpoint or data watchpoint, or halts itself.
@item Messages may be sent over ``debug message'' channels ... many
targets support such messages sent over JTAG,
for receipt by the person debugging or tools.
@item Loss of power ... some adapters can detect these events.
@item Resets not issued through JTAG ... such reset sources
can include button presses or other system hardware, sometimes
including the target itself (perhaps through a watchdog).
@item Debug instrumentation sometimes supports event triggering
such as ``trace buffer full'' (so it can quickly be emptied)
or other signals (to correlate with code behavior).
@end itemize

None of those events are signaled through standard JTAG signals.
However, most conventions for JTAG connectors include voltage
level and system reset (SRST) signal detection.
Some connectors also include instrumentation signals, which
can imply events when those signals are inputs.

In general, OpenOCD needs to periodically check for those events,
either by looking at the status of signals on the JTAG connector
or by sending synchronous ``tell me your status'' JTAG requests
to the various active targets.
There is a command to manage and monitor that polling,
which is normally done in the background.

@deffn Command poll [@option{on}|@option{off}]
Poll the current target for its current state.
(Also, @pxref{target curstate}.)
If that target is in debug mode, architecture
specific information about the current state is printed.
An optional parameter
allows background polling to be enabled and disabled.

You could use this from the TCL command shell, or
from GDB using @command{monitor poll} command.
@example
> poll
background polling: on
target state: halted
target halted in ARM state due to debug-request, \
               current mode: Supervisor
cpsr: 0x800000d3 pc: 0x11081bfc
MMU: disabled, D-Cache: disabled, I-Cache: enabled
>
@end example
@end deffn

@node Interface - Dongle Configuration
@chapter Interface - Dongle Configuration
@cindex config file, interface
@cindex interface config file

JTAG Adapters/Interfaces/Dongles are normally configured
through commands in an interface configuration
file which is sourced by your @file{openocd.cfg} file, or
through a command line @option{-f interface/....cfg} option.

@example
source [find interface/olimex-jtag-tiny.cfg]
@end example

These commands tell
OpenOCD what type of JTAG adapter you have, and how to talk to it.
A few cases are so simple that you only need to say what driver to use:

@example
# jlink interface
interface jlink
@end example

Most adapters need a bit more configuration than that.


@section Interface Configuration

The interface command tells OpenOCD what type of JTAG dongle you are
using. Depending on the type of dongle, you may need to have one or
more additional commands.

@deffn {Config Command} {interface} name
Use the interface driver @var{name} to connect to the
target.
@end deffn

@deffn Command {interface_list}
List the interface drivers that have been built into
the running copy of OpenOCD.
@end deffn

@deffn Command {jtag interface}
Returns the name of the interface driver being used.
@end deffn

@section Interface Drivers

Each of the interface drivers listed here must be explicitly
enabled when OpenOCD is configured, in order to be made
available at run time.

@deffn {Interface Driver} {amt_jtagaccel}
Amontec Chameleon in its JTAG Accelerator configuration,
connected to a PC's EPP mode parallel port.
This defines some driver-specific commands:

@deffn {Config Command} {parport_port} number
Specifies either the address of the I/O port (default: 0x378 for LPT1) or
the number of the @file{/dev/parport} device.
@end deffn

@deffn {Config Command} rtck [@option{enable}|@option{disable}]
Displays status of RTCK option.
Optionally sets that option first.
@end deffn
@end deffn

@deffn {Interface Driver} {arm-jtag-ew}
Olimex ARM-JTAG-EW USB adapter
This has one driver-specific command:

@deffn Command {armjtagew_info}
Logs some status
@end deffn
@end deffn

@deffn {Interface Driver} {at91rm9200}
Supports bitbanged JTAG from the local system,
presuming that system is an Atmel AT91rm9200
and a specific set of GPIOs is used.
@c command:	at91rm9200_device NAME
@c chooses among list of bit configs ... only one option
@end deffn

@deffn {Interface Driver} {dummy}
A dummy software-only driver for debugging.
@end deffn

@deffn {Interface Driver} {ep93xx}
Cirrus Logic EP93xx based single-board computer bit-banging (in development)
@end deffn

@deffn {Interface Driver} {ft2232}
FTDI FT2232 (USB) based devices over one of the userspace libraries.
These interfaces have several commands, used to configure the driver
before initializing the JTAG scan chain:

@deffn {Config Command} {ft2232_device_desc} description
Provides the USB device description (the @emph{iProduct string})
of the FTDI FT2232 device. If not
specified, the FTDI default value is used. This setting is only valid
if compiled with FTD2XX support.
@end deffn

@deffn {Config Command} {ft2232_serial} serial-number
Specifies the @var{serial-number} of the FTDI FT2232 device to use,
in case the vendor provides unique IDs and more than one FT2232 device
is connected to the host.
If not specified, serial numbers are not considered.
(Note that USB serial numbers can be arbitrary Unicode strings,
and are not restricted to containing only decimal digits.)
@end deffn

@deffn {Config Command} {ft2232_layout} name
Each vendor's FT2232 device can use different GPIO signals
to control output-enables, reset signals, and LEDs.
Currently valid layout @var{name} values include:
@itemize @minus
@item @b{axm0432_jtag} Axiom AXM-0432
@item @b{comstick} Hitex STR9 comstick
@item @b{cortino} Hitex Cortino JTAG interface
@item @b{evb_lm3s811} Luminary Micro EVB_LM3S811 as a JTAG interface,
either for the local Cortex-M3 (SRST only)
or in a passthrough mode (neither SRST nor TRST)
@item @b{luminary_icdi} Luminary In-Circuit Debug Interface (ICDI) Board
@item @b{flyswatter} Tin Can Tools Flyswatter
@item @b{icebear} ICEbear JTAG adapter from Section 5
@item @b{jtagkey} Amontec JTAGkey and JTAGkey-Tiny (and compatibles)
@item @b{jtagkey2} Amontec JTAGkey2 (and compatibles)
@item @b{m5960} American Microsystems M5960
@item @b{olimex-jtag} Olimex ARM-USB-OCD and ARM-USB-Tiny
@item @b{oocdlink} OOCDLink
@c oocdlink ~= jtagkey_prototype_v1
@item @b{sheevaplug} Marvell Sheevaplug development kit
@item @b{signalyzer} Xverve Signalyzer
@item @b{stm32stick} Hitex STM32 Performance Stick
@item @b{turtelizer2} egnite Software turtelizer2
@item @b{usbjtag} "USBJTAG-1" layout described in the OpenOCD diploma thesis
@end itemize
@end deffn

@deffn {Config Command} {ft2232_vid_pid} [vid pid]+
The vendor ID and product ID of the FTDI FT2232 device. If not specified, the FTDI
default values are used.
Currently, up to eight [@var{vid}, @var{pid}] pairs may be given, e.g.
@example
ft2232_vid_pid 0x0403 0xcff8 0x15ba 0x0003
@end example
@end deffn

@deffn {Config Command} {ft2232_latency} ms
On some systems using FT2232 based JTAG interfaces the FT_Read function call in
ft2232_read() fails to return the expected number of bytes. This can be caused by
USB communication delays and has proved hard to reproduce and debug. Setting the
FT2232 latency timer to a larger value increases delays for short USB packets but it
also reduces the risk of timeouts before receiving the expected number of bytes.
The OpenOCD default value is 2 and for some systems a value of 10 has proved useful.
@end deffn

For example, the interface config file for a
Turtelizer JTAG Adapter looks something like this:

@example
interface ft2232
ft2232_device_desc "Turtelizer JTAG/RS232 Adapter"
ft2232_layout turtelizer2
ft2232_vid_pid 0x0403 0xbdc8
@end example
@end deffn

@deffn {Interface Driver} {gw16012}
Gateworks GW16012 JTAG programmer.
This has one driver-specific command:

@deffn {Config Command} {parport_port} number
Specifies either the address of the I/O port (default: 0x378 for LPT1) or
the number of the @file{/dev/parport} device.
@end deffn
@end deffn

@deffn {Interface Driver} {jlink}
Segger jlink USB adapter
@c command:	jlink_info
@c     dumps status
@c command:	jlink_hw_jtag (2|3)
@c     sets version 2 or 3
@end deffn

@deffn {Interface Driver} {parport}
Supports PC parallel port bit-banging cables:
Wigglers, PLD download cable, and more.
These interfaces have several commands, used to configure the driver
before initializing the JTAG scan chain:

@deffn {Config Command} {parport_cable} name
The layout of the parallel port cable used to connect to the target.
Currently valid cable @var{name} values include:

@itemize @minus
@item @b{altium} Altium Universal JTAG cable.
@item @b{arm-jtag} Same as original wiggler except SRST and
TRST connections reversed and TRST is also inverted.
@item @b{chameleon} The Amontec Chameleon's CPLD when operated
in configuration mode. This is only used to
program the Chameleon itself, not a connected target.
@item @b{dlc5} The Xilinx Parallel cable III.
@item @b{flashlink} The ST Parallel cable.
@item @b{lattice} Lattice ispDOWNLOAD Cable
@item @b{old_amt_wiggler} The Wiggler configuration that comes with
some versions of
Amontec's Chameleon Programmer. The new version available from
the website uses the original Wiggler layout ('@var{wiggler}')
@item @b{triton} The parallel port adapter found on the
``Karo Triton 1 Development Board''.
This is also the layout used by the HollyGates design
(see @uref{http://www.lartmaker.nl/projects/jtag/}).
@item @b{wiggler} The original Wiggler layout, also supported by
several clones, such as the Olimex ARM-JTAG
@item @b{wiggler2} Same as original wiggler except an led is fitted on D5.
@item @b{wiggler_ntrst_inverted} Same as original wiggler except TRST is inverted.
@end itemize
@end deffn

@deffn {Config Command} {parport_port} number
Either the address of the I/O port (default: 0x378 for LPT1) or the number of
the @file{/dev/parport} device

When using PPDEV to access the parallel port, use the number of the parallel port:
@option{parport_port 0} (the default). If @option{parport_port 0x378} is specified
you may encounter a problem.
@end deffn

@deffn Command {parport_toggling_time} [nanoseconds]
Displays how many nanoseconds the hardware needs to toggle TCK;
the parport driver uses this value to obey the
@command{jtag_khz} configuration.
When the optional @var{nanoseconds} parameter is given,
that setting is changed before displaying the current value.

The default setting should work reasonably well on commodity PC hardware.
However, you may want to calibrate for your specific hardware.
@quotation Tip
To measure the toggling time with a logic analyzer or a digital storage
oscilloscope, follow the procedure below:
@example
> parport_toggling_time 1000
> jtag_khz 500
@end example
This sets the maximum JTAG clock speed of the hardware, but
the actual speed probably deviates from the requested 500 kHz.
Now, measure the time between the two closest spaced TCK transitions.
You can use @command{runtest 1000} or something similar to generate a
large set of samples.
Update the setting to match your measurement:
@example
> parport_toggling_time <measured nanoseconds>
@end example
Now the clock speed will be a better match for @command{jtag_khz rate}
commands given in OpenOCD scripts and event handlers.

You can do something similar with many digital multimeters, but note
that you'll probably need to run the clock continuously for several
seconds before it decides what clock rate to show.  Adjust the
toggling time up or down until the measured clock rate is a good
match for the jtag_khz rate you specified; be conservative.
@end quotation
@end deffn

@deffn {Config Command} {parport_write_on_exit} (on|off)
This will configure the parallel driver to write a known
cable-specific value to the parallel interface on exiting OpenOCD
@end deffn

For example, the interface configuration file for a
classic ``Wiggler'' cable might look something like this:

@example
interface parport
parport_port 0xc8b8
parport_cable wiggler
@end example
@end deffn

@deffn {Interface Driver} {presto}
ASIX PRESTO USB JTAG programmer.
@c command:	presto_serial str
@c     sets serial number
@end deffn

@deffn {Interface Driver} {rlink}
Raisonance RLink USB adapter
@end deffn

@deffn {Interface Driver} {usbprog}
usbprog is a freely programmable USB adapter.
@end deffn

@deffn {Interface Driver} {vsllink}
vsllink is part of Versaloon which is a versatile USB programmer.

@quotation Note
This defines quite a few driver-specific commands,
which are not currently documented here.
@end quotation
@end deffn

@deffn {Interface Driver} {ZY1000}
This is the Zylin ZY1000 JTAG debugger.

@quotation Note
This defines some driver-specific commands,
which are not currently documented here.
@end quotation

@deffn Command power [@option{on}|@option{off}]
Turn power switch to target on/off.
No arguments: print status.
@end deffn

@end deffn

@anchor{JTAG Speed}
@section JTAG Speed
JTAG clock setup is part of system setup.
It @emph{does not belong with interface setup} since any interface
only knows a few of the constraints for the JTAG clock speed.
Sometimes the JTAG speed is
changed during the target initialization process: (1) slow at
reset, (2) program the CPU clocks, (3) run fast.
Both the "slow" and "fast" clock rates are functions of the
oscillators used, the chip, the board design, and sometimes
power management software that may be active.

The speed used during reset, and the scan chain verification which
follows reset, can be adjusted using a @code{reset-start}
target event handler.
It can then be reconfigured to a faster speed by a
@code{reset-init} target event handler after it reprograms those
CPU clocks, or manually (if something else, such as a boot loader,
sets up those clocks).
@xref{Target Events}.
When the initial low JTAG speed is a chip characteristic, perhaps
because of a required oscillator speed, provide such a handler
in the target config file.
When that speed is a function of a board-specific characteristic
such as which speed oscillator is used, it belongs in the board
config file instead.
In both cases it's safest to also set the initial JTAG clock rate
to that same slow speed, so that OpenOCD never starts up using a
clock speed that's faster than the scan chain can support.

@example
jtag_rclk 3000
$_TARGET.cpu configure -event reset-start @{ jtag_rclk 3000 @}
@end example

If your system supports adaptive clocking (RTCK), configuring
JTAG to use that is probably the most robust approach.
However, it introduces delays to synchronize clocks; so it
may not be the fastest solution.

@b{NOTE:} Script writers should consider using @command{jtag_rclk}
instead of @command{jtag_khz}.

@deffn {Command} jtag_khz max_speed_kHz
A non-zero speed is in KHZ. Hence: 3000 is 3mhz.
JTAG interfaces usually support a limited number of
speeds.  The speed actually used won't be faster
than the speed specified.

Chip data sheets generally include a top JTAG clock rate.
The actual rate is often a function of a CPU core clock,
and is normally less than that peak rate.
For example, most ARM cores accept at most one sixth of the CPU clock.

Speed 0 (khz) selects RTCK method.
@xref{FAQ RTCK}.
If your system uses RTCK, you won't need to change the
JTAG clocking after setup.
Not all interfaces, boards, or targets support ``rtck''.
If the interface device can not
support it, an error is returned when you try to use RTCK.
@end deffn

@defun jtag_rclk fallback_speed_kHz
@cindex adaptive clocking
@cindex RTCK
This Tcl proc (defined in @file{startup.tcl}) attempts to enable RTCK/RCLK.
If that fails (maybe the interface, board, or target doesn't
support it), falls back to the specified frequency.
@example
# Fall back to 3mhz if RTCK is not supported
jtag_rclk 3000
@end example
@end defun

@node Reset Configuration
@chapter Reset Configuration
@cindex Reset Configuration

Every system configuration may require a different reset
configuration. This can also be quite confusing.
Resets also interact with @var{reset-init} event handlers,
which do things like setting up clocks and DRAM, and
JTAG clock rates.  (@xref{JTAG Speed}.)
They can also interact with JTAG routers.
Please see the various board files for examples.

@quotation Note
To maintainers and integrators:
Reset configuration touches several things at once.
Normally the board configuration file
should define it and assume that the JTAG adapter supports
everything that's wired up to the board's JTAG connector.

However, the target configuration file could also make note
of something the silicon vendor has done inside the chip,
which will be true for most (or all) boards using that chip.
And when the JTAG adapter doesn't support everything, the
user configuration file will need to override parts of
the reset configuration provided by other files.
@end quotation

@section Types of Reset

There are many kinds of reset possible through JTAG, but
they may not all work with a given board and adapter.
That's part of why reset configuration can be error prone.

@itemize @bullet
@item
@emph{System Reset} ... the @emph{SRST} hardware signal
resets all chips connected to the JTAG adapter, such as processors,
power management chips, and I/O controllers.  Normally resets triggered
with this signal behave exactly like pressing a RESET button.
@item
@emph{JTAG TAP Reset} ... the @emph{TRST} hardware signal resets
just the TAP controllers connected to the JTAG adapter.
Such resets should not be visible to the rest of the system; resetting a
device's the TAP controller just puts that controller into a known state.
@item
@emph{Emulation Reset} ... many devices can be reset through JTAG
commands.  These resets are often distinguishable from system
resets, either explicitly (a "reset reason" register says so)
or implicitly (not all parts of the chip get reset).
@item
@emph{Other Resets} ... system-on-chip devices often support
several other types of reset.
You may need to arrange that a watchdog timer stops
while debugging, preventing a watchdog reset.
There may be individual module resets.
@end itemize

In the best case, OpenOCD can hold SRST, then reset
the TAPs via TRST and send commands through JTAG to halt the
CPU at the reset vector before the 1st instruction is executed.
Then when it finally releases the SRST signal, the system is
halted under debugger control before any code has executed.
This is the behavior required to support the @command{reset halt}
and @command{reset init} commands; after @command{reset init} a
board-specific script might do things like setting up DRAM.
(@xref{Reset Command}.)

@anchor{SRST and TRST Issues}
@section SRST and TRST Issues

Because SRST and TRST are hardware signals, they can have a
variety of system-specific constraints.  Some of the most
common issues are:

@itemize @bullet

@item @emph{Signal not available} ... Some boards don't wire
SRST or TRST to the JTAG connector.  Some JTAG adapters don't
support such signals even if they are wired up.
Use the @command{reset_config} @var{signals} options to say
when either of those signals is not connected.
When SRST is not available, your code might not be able to rely
on controllers having been fully reset during code startup.
Missing TRST is not a problem, since JTAG level resets can
be triggered using with TMS signaling.

@item @emph{Signals shorted} ... Sometimes a chip, board, or
adapter will connect SRST to TRST, instead of keeping them separate.
Use the @command{reset_config} @var{combination} options to say
when those signals aren't properly independent.

@item @emph{Timing} ... Reset circuitry like a resistor/capacitor
delay circuit, reset supervisor, or on-chip features can extend
the effect of a JTAG adapter's reset for some time after the adapter
stops issuing the reset.  For example, there may be chip or board
requirements that all reset pulses last for at least a
certain amount of time; and reset buttons commonly have
hardware debouncing.
Use the @command{jtag_nsrst_delay} and @command{jtag_ntrst_delay}
commands to say when extra delays are needed.

@item @emph{Drive type} ... Reset lines often have a pullup
resistor, letting the JTAG interface treat them as open-drain
signals.  But that's not a requirement, so the adapter may need
to use push/pull output drivers.
Also, with weak pullups it may be advisable to drive
signals to both levels (push/pull) to minimize rise times.
Use the @command{reset_config} @var{trst_type} and
@var{srst_type} parameters to say how to drive reset signals.

@item @emph{Special initialization} ...  Targets sometimes need
special JTAG initialization sequences to handle chip-specific
issues (not limited to errata).
For example, certain JTAG commands might need to be issued while
the system as a whole is in a reset state (SRST active)
but the JTAG scan chain is usable (TRST inactive).
Many systems treat combined assertion of SRST and TRST as a
trigger for a harder reset than SRST alone.
Such custom reset handling is discussed later in this chapter.
@end itemize

There can also be other issues.
Some devices don't fully conform to the JTAG specifications.
Trivial system-specific differences are common, such as
SRST and TRST using slightly different names.
There are also vendors who distribute key JTAG documentation for
their chips only to developers who have signed a Non-Disclosure
Agreement (NDA).

Sometimes there are chip-specific extensions like a requirement to use
the normally-optional TRST signal (precluding use of JTAG adapters which
don't pass TRST through), or needing extra steps to complete a TAP reset.

In short, SRST and especially TRST handling may be very finicky,
needing to cope with both architecture and board specific constraints.

@section Commands for Handling Resets

@deffn {Command} jtag_nsrst_assert_width milliseconds
Minimum amount of time (in milliseconds) OpenOCD should wait
after asserting nSRST (active-low system reset) before
allowing it to be deasserted.
@end deffn

@deffn {Command} jtag_nsrst_delay milliseconds
How long (in milliseconds) OpenOCD should wait after deasserting
nSRST (active-low system reset) before starting new JTAG operations.
When a board has a reset button connected to SRST line it will
probably have hardware debouncing, implying you should use this.
@end deffn

@deffn {Command} jtag_ntrst_assert_width milliseconds
Minimum amount of time (in milliseconds) OpenOCD should wait
after asserting nTRST (active-low JTAG TAP reset) before
allowing it to be deasserted.
@end deffn

@deffn {Command} jtag_ntrst_delay milliseconds
How long (in milliseconds) OpenOCD should wait after deasserting
nTRST (active-low JTAG TAP reset) before starting new JTAG operations.
@end deffn

@deffn {Command} reset_config mode_flag ...
This command displays or modifies the reset configuration
of your combination of JTAG board and target in target
configuration scripts.

Information earlier in this section describes the kind of problems
the command is intended to address (@pxref{SRST and TRST Issues}).
As a rule this command belongs only in board config files,
describing issues like @emph{board doesn't connect TRST};
or in user config files, addressing limitations derived
from a particular combination of interface and board.
(An unlikely example would be using a TRST-only adapter
with a board that only wires up SRST.)

The @var{mode_flag} options can be specified in any order, but only one
of each type -- @var{signals}, @var{combination},
@var{gates},
@var{trst_type},
and @var{srst_type} -- may be specified at a time.
If you don't provide a new value for a given type, its previous
value (perhaps the default) is unchanged.
For example, this means that you don't need to say anything at all about
TRST just to declare that if the JTAG adapter should want to drive SRST,
it must explicitly be driven high (@option{srst_push_pull}).

@itemize
@item
@var{signals} can specify which of the reset signals are connected.
For example, If the JTAG interface provides SRST, but the board doesn't
connect that signal properly, then OpenOCD can't use it.
Possible values are @option{none} (the default), @option{trst_only},
@option{srst_only} and @option{trst_and_srst}.

@quotation Tip
If your board provides SRST and/or TRST through the JTAG connector,
you must declare that so those signals can be used.
@end quotation

@item
The @var{combination} is an optional value specifying broken reset
signal implementations.
The default behaviour if no option given is @option{separate},
indicating everything behaves normally.
@option{srst_pulls_trst} states that the
test logic is reset together with the reset of the system (e.g. Philips
LPC2000, "broken" board layout), @option{trst_pulls_srst} says that
the system is reset together with the test logic (only hypothetical, I
haven't seen hardware with such a bug, and can be worked around).
@option{combined} implies both @option{srst_pulls_trst} and
@option{trst_pulls_srst}.

@item
The @var{gates} tokens control flags that describe some cases where
JTAG may be unvailable during reset.
@option{srst_gates_jtag} (default)
indicates that asserting SRST gates the
JTAG clock. This means that no communication can happen on JTAG
while SRST is asserted.
Its converse is @option{srst_nogate}, indicating that JTAG commands
can safely be issued while SRST is active.
@end itemize

The optional @var{trst_type} and @var{srst_type} parameters allow the
driver mode of each reset line to be specified.  These values only affect
JTAG interfaces with support for different driver modes, like the Amontec
JTAGkey and JTAG Accelerator.  Also, they are necessarily ignored if the
relevant signal (TRST or SRST) is not connected.

@itemize
@item
Possible @var{trst_type} driver modes for the test reset signal (TRST)
are the default @option{trst_push_pull}, and @option{trst_open_drain}.
Most boards connect this signal to a pulldown, so the JTAG TAPs
never leave reset unless they are hooked up to a JTAG adapter.

@item
Possible @var{srst_type} driver modes for the system reset signal (SRST)
are the default @option{srst_open_drain}, and @option{srst_push_pull}.
Most boards connect this signal to a pullup, and allow the
signal to be pulled low by various events including system
powerup and pressing a reset button.
@end itemize
@end deffn

@section Custom Reset Handling
@cindex events

OpenOCD has several ways to help support the various reset
mechanisms provided by chip and board vendors.
The commands shown in the previous section give standard parameters.
There are also @emph{event handlers} associated with TAPs or Targets.
Those handlers are Tcl procedures you can provide, which are invoked
at particular points in the reset sequence.

@emph{When SRST is not an option} you must set
up a @code{reset-assert} event handler for your target.
For example, some JTAG adapters don't include the SRST signal;
and some boards have multiple targets, and you won't always
want to reset everything at once.

After configuring those mechanisms, you might still
find your board doesn't start up or reset correctly.
For example, maybe it needs a slightly different sequence
of SRST and/or TRST manipulations, because of quirks that
the @command{reset_config} mechanism doesn't address;
or asserting both might trigger a stronger reset, which
needs special attention.

Experiment with lower level operations, such as @command{jtag_reset}
and the @command{jtag arp_*} operations shown here,
to find a sequence of operations that works.
@xref{JTAG Commands}.
When you find a working sequence, it can be used to override
@command{jtag_init}, which fires during OpenOCD startup
(@pxref{Configuration Stage});
or @command{init_reset}, which fires during reset processing.

You might also want to provide some project-specific reset
schemes.  For example, on a multi-target board the standard
@command{reset} command would reset all targets, but you
may need the ability to reset only one target at time and
thus want to avoid using the board-wide SRST signal.

@deffn {Overridable Procedure} init_reset mode
This is invoked near the beginning of the @command{reset} command,
usually to provide as much of a cold (power-up) reset as practical.
By default it is also invoked from @command{jtag_init} if
the scan chain does not respond to pure JTAG operations.
The @var{mode} parameter is the parameter given to the
low level reset command (@option{halt},
@option{init}, or @option{run}), @option{setup},
or potentially some other value.

The default implementation just invokes @command{jtag arp_init-reset}.
Replacements will normally build on low level JTAG
operations such as @command{jtag_reset}.
Operations here must not address individual TAPs
(or their associated targets)
until the JTAG scan chain has first been verified to work.

Implementations must have verified the JTAG scan chain before
they return.
This is done by calling @command{jtag arp_init}
(or @command{jtag arp_init-reset}).
@end deffn

@deffn Command {jtag arp_init}
This validates the scan chain using just the four
standard JTAG signals (TMS, TCK, TDI, TDO).
It starts by issuing a JTAG-only reset.
Then it performs checks to verify that the scan chain configuration
matches the TAPs it can observe.
Those checks include checking IDCODE values for each active TAP,
and verifying the length of their instruction registers using
TAP @code{-ircapture} and @code{-irmask} values.
If these tests all pass, TAP @code{setup} events are
issued to all TAPs with handlers for that event.
@end deffn

@deffn Command {jtag arp_init-reset}
This uses TRST and SRST to try resetting
everything on the JTAG scan chain
(and anything else connected to SRST).
It then invokes the logic of @command{jtag arp_init}.
@end deffn


@node TAP Declaration
@chapter TAP Declaration
@cindex TAP declaration
@cindex TAP configuration

@emph{Test Access Ports} (TAPs) are the core of JTAG.
TAPs serve many roles, including:

@itemize @bullet
@item @b{Debug Target} A CPU TAP can be used as a GDB debug target
@item @b{Flash Programing} Some chips program the flash directly via JTAG.
Others do it indirectly, making a CPU do it.
@item @b{Program Download} Using the same CPU support GDB uses,
you can initialize a DRAM controller, download code to DRAM, and then
start running that code.
@item @b{Boundary Scan} Most chips support boundary scan, which
helps test for board assembly problems like solder bridges
and missing connections
@end itemize

OpenOCD must know about the active TAPs on your board(s).
Setting up the TAPs is the core task of your configuration files.
Once those TAPs are set up, you can pass their names to code
which sets up CPUs and exports them as GDB targets,
probes flash memory, performs low-level JTAG operations, and more.

@section Scan Chains
@cindex scan chain

TAPs are part of a hardware @dfn{scan chain},
which is daisy chain of TAPs.
They also need to be added to
OpenOCD's software mirror of that hardware list,
giving each member a name and associating other data with it.
Simple scan chains, with a single TAP, are common in
systems with a single microcontroller or microprocessor.
More complex chips may have several TAPs internally.
Very complex scan chains might have a dozen or more TAPs:
several in one chip, more in the next, and connecting
to other boards with their own chips and TAPs.

You can display the list with the @command{scan_chain} command.
(Don't confuse this with the list displayed by the @command{targets}
command, presented in the next chapter.
That only displays TAPs for CPUs which are configured as
debugging targets.)
Here's what the scan chain might look like for a chip more than one TAP:

@verbatim
   TapName            Enabled IdCode     Expected   IrLen IrCap IrMask
-- ------------------ ------- ---------- ---------- ----- ----- ------
 0 omap5912.dsp          Y    0x03df1d81 0x03df1d81    38 0x01  0x03
 1 omap5912.arm          Y    0x0692602f 0x0692602f     4 0x01  0x0f
 2 omap5912.unknown      Y    0x00000000 0x00000000     8 0x01  0x03
@end verbatim

OpenOCD can detect some of that information, but not all
of it.  @xref{Autoprobing}.
Unfortunately those TAPs can't always be autoconfigured,
because not all devices provide good support for that.
JTAG doesn't require supporting IDCODE instructions, and
chips with JTAG routers may not link TAPs into the chain
until they are told to do so.

The configuration mechanism currently supported by OpenOCD
requires explicit configuration of all TAP devices using
@command{jtag newtap} commands, as detailed later in this chapter.
A command like this would declare one tap and name it @code{chip1.cpu}:

@example
jtag newtap chip1 cpu -irlen 4 -expected-id 0x3ba00477
@end example

Each target configuration file lists the TAPs provided
by a given chip.
Board configuration files combine all the targets on a board,
and so forth.
Note that @emph{the order in which TAPs are declared is very important.}
It must match the order in the JTAG scan chain, both inside
a single chip and between them.
@xref{FAQ TAP Order}.

For example, the ST Microsystems STR912 chip has
three separate TAPs@footnote{See the ST
document titled: @emph{STR91xFAxxx, Section 3.15 Jtag Interface, Page:
28/102, Figure 3: JTAG chaining inside the STR91xFA}.
@url{http://eu.st.com/stonline/products/literature/ds/13495.pdf}}.
To configure those taps, @file{target/str912.cfg}
includes commands something like this:

@example
jtag newtap str912 flash ... params ...
jtag newtap str912 cpu ... params ...
jtag newtap str912 bs ... params ...
@end example

Actual config files use a variable instead of literals like
@option{str912}, to support more than one chip of each type.
@xref{Config File Guidelines}.

@deffn Command {jtag names}
Returns the names of all current TAPs in the scan chain.
Use @command{jtag cget} or @command{jtag tapisenabled}
to examine attributes and state of each TAP.
@example
foreach t [jtag names] @{
    puts [format "TAP: %s\n" $t]
@}
@end example
@end deffn

@deffn Command {scan_chain}
Displays the TAPs in the scan chain configuration,
and their status.
The set of TAPs listed by this command is fixed by
exiting the OpenOCD configuration stage,
but systems with a JTAG router can
enable or disable TAPs dynamically.
@end deffn

@c FIXME!  "jtag cget" should be able to return all TAP
@c attributes, like "$target_name cget" does for targets.

@c Probably want "jtag eventlist", and a "tap-reset" event
@c (on entry to RESET state).

@section TAP Names
@cindex dotted name

When TAP objects are declared with @command{jtag newtap},
a @dfn{dotted.name} is created for the TAP, combining the
name of a module (usually a chip) and a label for the TAP.
For example: @code{xilinx.tap}, @code{str912.flash},
@code{omap3530.jrc}, @code{dm6446.dsp}, or @code{stm32.cpu}.
Many other commands use that dotted.name to manipulate or
refer to the TAP.  For example, CPU configuration uses the
name, as does declaration of NAND or NOR flash banks.

The components of a dotted name should follow ``C'' symbol
name rules:  start with an alphabetic character, then numbers
and underscores are OK; while others (including dots!) are not.

@quotation Tip
In older code, JTAG TAPs were numbered from 0..N.
This feature is still present.
However its use is highly discouraged, and
should not be relied on; it will be removed by mid-2010.
Update all of your scripts to use TAP names rather than numbers,
by paying attention to the runtime warnings they trigger.
Using TAP numbers in target configuration scripts prevents
reusing those scripts on boards with multiple targets.
@end quotation

@section TAP Declaration Commands

@c shouldn't this be(come) a {Config Command}?
@anchor{jtag newtap}
@deffn Command {jtag newtap} chipname tapname configparams...
Declares a new TAP with the dotted name @var{chipname}.@var{tapname},
and configured according to the various @var{configparams}.

The @var{chipname} is a symbolic name for the chip.
Conventionally target config files use @code{$_CHIPNAME},
defaulting to the model name given by the chip vendor but
overridable.

@cindex TAP naming convention
The @var{tapname} reflects the role of that TAP,
and should follow this convention:

@itemize @bullet
@item @code{bs} -- For boundary scan if this is a seperate TAP;
@item @code{cpu} -- The main CPU of the chip, alternatively
@code{arm} and @code{dsp} on chips with both ARM and DSP CPUs,
@code{arm1} and @code{arm2} on chips two ARMs, and so forth;
@item @code{etb} -- For an embedded trace buffer (example: an ARM ETB11);
@item @code{flash} -- If the chip has a flash TAP, like the str912;
@item @code{jrc} -- For JTAG route controller (example: the ICEpick modules
on many Texas Instruments chips, like the OMAP3530 on Beagleboards);
@item @code{tap} -- Should be used only FPGA or CPLD like devices
with a single TAP;
@item @code{unknownN} -- If you have no idea what the TAP is for (N is a number);
@item @emph{when in doubt} -- Use the chip maker's name in their data sheet.
For example, the Freescale IMX31 has a SDMA (Smart DMA) with
a JTAG TAP; that TAP should be named @code{sdma}.
@end itemize

Every TAP requires at least the following @var{configparams}:

@itemize @bullet
@item @code{-irlen} @var{NUMBER}
@*The length in bits of the
instruction register, such as 4 or 5 bits.
@end itemize

A TAP may also provide optional @var{configparams}:

@itemize @bullet
@item @code{-disable} (or @code{-enable})
@*Use the @code{-disable} parameter to flag a TAP which is not
linked in to the scan chain after a reset using either TRST
or the JTAG state machine's @sc{reset} state.
You may use @code{-enable} to highlight the default state
(the TAP is linked in).
@xref{Enabling and Disabling TAPs}.
@item @code{-expected-id} @var{number}
@*A non-zero @var{number} represents a 32-bit IDCODE
which you expect to find when the scan chain is examined.
These codes are not required by all JTAG devices.
@emph{Repeat the option} as many times as required if more than one
ID code could appear (for example, multiple versions).
Specify @var{number} as zero to suppress warnings about IDCODE
values that were found but not included in the list.

Provide this value if at all possible, since it lets OpenOCD
tell when the scan chain it sees isn't right.  These values
are provided in vendors' chip documentation, usually a technical
reference manual.  Sometimes you may need to probe the JTAG
hardware to find these values.
@xref{Autoprobing}.
@item @code{-ignore-version}
@*Specify this to ignore the JTAG version field in the @code{-expected-id}
option.  When vendors put out multiple versions of a chip, or use the same
JTAG-level ID for several largely-compatible chips, it may be more practical
to ignore the version field than to update config files to handle all of
the various chip IDs.
@item @code{-ircapture} @var{NUMBER}
@*The bit pattern loaded by the TAP into the JTAG shift register
on entry to the @sc{ircapture} state, such as 0x01.
JTAG requires the two LSBs of this value to be 01.
By default, @code{-ircapture} and @code{-irmask} are set
up to verify that two-bit value.  You may provide
additional bits, if you know them, or indicate that
a TAP doesn't conform to the JTAG specification.
@item @code{-irmask} @var{NUMBER}
@*A mask used with @code{-ircapture}
to verify that instruction scans work correctly.
Such scans are not used by OpenOCD except to verify that
there seems to be no problems with JTAG scan chain operations.
@end itemize
@end deffn

@section Other TAP commands

@deffn Command {jtag cget} dotted.name @option{-event} name
@deffnx Command {jtag configure} dotted.name @option{-event} name string
At this writing this TAP attribute
mechanism is used only for event handling.
(It is not a direct analogue of the @code{cget}/@code{configure}
mechanism for debugger targets.)
See the next section for information about the available events.

The @code{configure} subcommand assigns an event handler,
a TCL string which is evaluated when the event is triggered.
The @code{cget} subcommand returns that handler.
@end deffn

@anchor{TAP Events}
@section TAP Events
@cindex events
@cindex TAP events

OpenOCD includes two event mechanisms.
The one presented here applies to all JTAG TAPs.
The other applies to debugger targets,
which are associated with certain TAPs.

The TAP events currently defined are:

@itemize @bullet
@item @b{post-reset}
@* The TAP has just completed a JTAG reset.
The tap may still be in the JTAG @sc{reset} state.
Handlers for these events might perform initialization sequences
such as issuing TCK cycles, TMS sequences to ensure
exit from the ARM SWD mode, and more.

Because the scan chain has not yet been verified, handlers for these events
@emph{should not issue commands which scan the JTAG IR or DR registers}
of any particular target.
@b{NOTE:} As this is written (September 2009), nothing prevents such access.
@item @b{setup}
@* The scan chain has been reset and verified.
This handler may enable TAPs as needed.
@item @b{tap-disable}
@* The TAP needs to be disabled.  This handler should
implement @command{jtag tapdisable}
by issuing the relevant JTAG commands.
@item @b{tap-enable}
@* The TAP needs to be enabled.  This handler should
implement @command{jtag tapenable}
by issuing the relevant JTAG commands.
@end itemize

If you need some action after each JTAG reset, which isn't actually
specific to any TAP (since you can't yet trust the scan chain's
contents to be accurate), you might:

@example
jtag configure CHIP.jrc -event post-reset @{
  echo "JTAG Reset done"
  ... non-scan jtag operations to be done after reset
@}
@end example


@anchor{Enabling and Disabling TAPs}
@section Enabling and Disabling TAPs
@cindex JTAG Route Controller
@cindex jrc

In some systems, a @dfn{JTAG Route Controller} (JRC)
is used to enable and/or disable specific JTAG TAPs.
Many ARM based chips from Texas Instruments include
an ``ICEpick'' module, which is a JRC.
Such chips include DaVinci and OMAP3 processors.

A given TAP may not be visible until the JRC has been
told to link it into the scan chain; and if the JRC
has been told to unlink that TAP, it will no longer
be visible.
Such routers address problems that JTAG ``bypass mode''
ignores, such as:

@itemize
@item The scan chain can only go as fast as its slowest TAP.
@item Having many TAPs slows instruction scans, since all
TAPs receive new instructions.
@item TAPs in the scan chain must be powered up, which wastes
power and prevents debugging some power management mechanisms.
@end itemize

The IEEE 1149.1 JTAG standard has no concept of a ``disabled'' tap,
as implied by the existence of JTAG routers.
However, the upcoming IEEE 1149.7 framework (layered on top of JTAG)
does include a kind of JTAG router functionality.

@c (a) currently the event handlers don't seem to be able to
@c     fail in a way that could lead to no-change-of-state.

In OpenOCD, tap enabling/disabling is invoked by the Tcl commands
shown below, and is implemented using TAP event handlers.
So for example, when defining a TAP for a CPU connected to
a JTAG router, your @file{target.cfg} file
should define TAP event handlers using
code that looks something like this:

@example
jtag configure CHIP.cpu -event tap-enable @{
  ... jtag operations using CHIP.jrc
@}
jtag configure CHIP.cpu -event tap-disable @{
  ... jtag operations using CHIP.jrc
@}
@end example

Then you might want that CPU's TAP enabled almost all the time:

@example
jtag configure $CHIP.jrc -event setup "jtag tapenable $CHIP.cpu"
@end example

Note how that particular setup event handler declaration
uses quotes to evaluate @code{$CHIP} when the event is configured.
Using brackets @{ @} would cause it to be evaluated later,
at runtime, when it might have a different value.

@deffn Command {jtag tapdisable} dotted.name
If necessary, disables the tap
by sending it a @option{tap-disable} event.
Returns the string "1" if the tap
specified by @var{dotted.name} is enabled,
and "0" if it is disabled.
@end deffn

@deffn Command {jtag tapenable} dotted.name
If necessary, enables the tap
by sending it a @option{tap-enable} event.
Returns the string "1" if the tap
specified by @var{dotted.name} is enabled,
and "0" if it is disabled.
@end deffn

@deffn Command {jtag tapisenabled} dotted.name
Returns the string "1" if the tap
specified by @var{dotted.name} is enabled,
and "0" if it is disabled.

@quotation Note
Humans will find the @command{scan_chain} command more helpful
for querying the state of the JTAG taps.
@end quotation
@end deffn

@anchor{Autoprobing}
@section Autoprobing
@cindex autoprobe
@cindex JTAG autoprobe

TAP configuration is the first thing that needs to be done
after interface and reset configuration.  Sometimes it's
hard finding out what TAPs exist, or how they are identified.
Vendor documentation is not always easy to find and use.

To help you get past such problems, OpenOCD has a limited
@emph{autoprobing} ability to look at the scan chain, doing
a @dfn{blind interrogation} and then reporting the TAPs it finds.
To use this mechanism, start the OpenOCD server with only data
that configures your JTAG interface, and arranges to come up
with a slow clock (many devices don't support fast JTAG clocks
right when they come out of reset).

For example, your @file{openocd.cfg} file might have:

@example
source [find interface/olimex-arm-usb-tiny-h.cfg]
reset_config trst_and_srst
jtag_rclk 8
@end example

When you start the server without any TAPs configured, it will
attempt to autoconfigure the TAPs.  There are two parts to this:

@enumerate
@item @emph{TAP discovery} ...
After a JTAG reset (sometimes a system reset may be needed too),
each TAP's data registers will hold the contents of either the
IDCODE or BYPASS register.
If JTAG communication is working, OpenOCD will see each TAP,
and report what @option{-expected-id} to use with it.
@item @emph{IR Length discovery} ...
Unfortunately JTAG does not provide a reliable way to find out
the value of the @option{-irlen} parameter to use with a TAP
that is discovered.
If OpenOCD can discover the length of a TAP's instruction
register, it will report it.
Otherwise you may need to consult vendor documentation, such
as chip data sheets or BSDL files.
@end enumerate

In many cases your board will have a simple scan chain with just
a single device.  Here's what OpenOCD reported with one board
that's a bit more complex:

@example
clock speed 8 kHz
There are no enabled taps.  AUTO PROBING MIGHT NOT WORK!!
AUTO auto0.tap - use "jtag newtap auto0 tap -expected-id 0x2b900f0f ..."
AUTO auto1.tap - use "jtag newtap auto1 tap -expected-id 0x07926001 ..."
AUTO auto2.tap - use "jtag newtap auto2 tap -expected-id 0x0b73b02f ..."
AUTO auto0.tap - use "... -irlen 4"
AUTO auto1.tap - use "... -irlen 4"
AUTO auto2.tap - use "... -irlen 6"
no gdb ports allocated as no target has been specified
@end example

Given that information, you should be able to either find some existing
config files to use, or create your own.  If you create your own, you
would configure from the bottom up:  first a @file{target.cfg} file
with these TAPs, any targets associated with them, and any on-chip
resources; then a @file{board.cfg} with off-chip resources, clocking,
and so forth.

@node CPU Configuration
@chapter CPU Configuration
@cindex GDB target

This chapter discusses how to set up GDB debug targets for CPUs.
You can also access these targets without GDB
(@pxref{Architecture and Core Commands},
and @ref{Target State handling}) and
through various kinds of NAND and NOR flash commands.
If you have multiple CPUs you can have multiple such targets.

We'll start by looking at how to examine the targets you have,
then look at how to add one more target and how to configure it.

@section Target List
@cindex target, current
@cindex target, list

All targets that have been set up are part of a list,
where each member has a name.
That name should normally be the same as the TAP name.
You can display the list with the @command{targets}
(plural!) command.
This display often has only one CPU; here's what it might
look like with more than one:
@verbatim
    TargetName         Type       Endian TapName            State
--  ------------------ ---------- ------ ------------------ ------------
 0* at91rm9200.cpu     arm920t    little at91rm9200.cpu     running
 1  MyTarget           cortex_m3  little mychip.foo         tap-disabled
@end verbatim

One member of that list is the @dfn{current target}, which
is implicitly referenced by many commands.
It's the one marked with a @code{*} near the target name.
In particular, memory addresses often refer to the address
space seen by that current target.
Commands like @command{mdw} (memory display words)
and @command{flash erase_address} (erase NOR flash blocks)
are examples; and there are many more.

Several commands let you examine the list of targets:

@deffn Command {target count}
@emph{Note: target numbers are deprecated; don't use them.
They will be removed shortly after August 2010, including this command.
Iterate target using @command{target names}, not by counting.}

Returns the number of targets, @math{N}.
The highest numbered target is @math{N - 1}.
@example
set c [target count]
for @{ set x 0 @} @{ $x < $c @} @{ incr x @} @{
    # Assuming you have created this function
    print_target_details $x
@}
@end example
@end deffn

@deffn Command {target current}
Returns the name of the current target.
@end deffn

@deffn Command {target names}
Lists the names of all current targets in the list.
@example
foreach t [target names] @{
    puts [format "Target: %s\n" $t]
@}
@end example
@end deffn

@deffn Command {target number} number
@emph{Note: target numbers are deprecated; don't use them.
They will be removed shortly after August 2010, including this command.}

The list of targets is numbered starting at zero.
This command returns the name of the target at index @var{number}.
@example
set thename [target number $x]
puts [format "Target %d is: %s\n" $x $thename]
@end example
@end deffn

@c yep, "target list" would have been better.
@c plus maybe "target setdefault".

@deffn Command targets [name]
@emph{Note: the name of this command is plural.  Other target
command names are singular.}

With no parameter, this command displays a table of all known
targets in a user friendly form.

With a parameter, this command sets the current target to
the given target with the given @var{name}; this is
only relevant on boards which have more than one target.
@end deffn

@section Target CPU Types and Variants
@cindex target type
@cindex CPU type
@cindex CPU variant

Each target has a @dfn{CPU type}, as shown in the output of
the @command{targets} command.  You need to specify that type
when calling @command{target create}.
The CPU type indicates more than just the instruction set.
It also indicates how that instruction set is implemented,
what kind of debug support it integrates,
whether it has an MMU (and if so, what kind),
what core-specific commands may be available
(@pxref{Architecture and Core Commands}),
and more.

For some CPU types, OpenOCD also defines @dfn{variants} which
indicate differences that affect their handling.
For example, a particular implementation bug might need to be
worked around in some chip versions.

It's easy to see what target types are supported,
since there's a command to list them.
However, there is currently no way to list what target variants
are supported (other than by reading the OpenOCD source code).

@anchor{target types}
@deffn Command {target types}
Lists all supported target types.
At this writing, the supported CPU types and variants are:

@itemize @bullet
@item @code{arm11} -- this is a generation of ARMv6 cores
@item @code{arm720t} -- this is an ARMv4 core with an MMU
@item @code{arm7tdmi} -- this is an ARMv4 core
@item @code{arm920t} -- this is an ARMv5 core with an MMU
@item @code{arm926ejs} -- this is an ARMv5 core with an MMU
@item @code{arm966e} -- this is an ARMv5 core
@item @code{arm9tdmi} -- this is an ARMv4 core
@item @code{avr} -- implements Atmel's 8-bit AVR instruction set.
(Support for this is preliminary and incomplete.)
@item @code{cortex_a8} -- this is an ARMv7 core with an MMU
@item @code{cortex_m3} -- this is an ARMv7 core, supporting only the
compact Thumb2 instruction set.  It supports one variant:
@itemize @minus
@item @code{lm3s} ... Use this when debugging older Stellaris LM3S targets.
This will cause OpenOCD to use a software reset rather than asserting
SRST, to avoid a issue with clearing the debug registers.
This is fixed in Fury Rev B, DustDevil Rev B, Tempest; these revisions will
be detected and the normal reset behaviour used.
@end itemize
@item @code{dragonite} -- resembles arm966e
@item @code{dsp563xx} -- implements Freescale's 24-bit DSP.
(Support for this is still incomplete.)
@item @code{fa526} -- resembles arm920 (w/o Thumb)
@item @code{feroceon} -- resembles arm926
@item @code{mips_m4k} -- a MIPS core.  This supports one variant:
@itemize @minus
@item @code{ejtag_srst} ... Use this when debugging targets that do not
provide a functional SRST line on the EJTAG connector.  This causes
OpenOCD to instead use an EJTAG software reset command to reset the
processor.
You still need to enable @option{srst} on the @command{reset_config}
command to enable OpenOCD hardware reset functionality.
@end itemize
@item @code{xscale} -- this is actually an architecture,
not a CPU type.  It is based on the ARMv5 architecture.
There are several variants defined:
@itemize @minus
@item @code{ixp42x}, @code{ixp45x}, @code{ixp46x},
@code{pxa27x} ... instruction register length is 7 bits
@item @code{pxa250}, @code{pxa255},
@code{pxa26x} ... instruction register length is 5 bits
@item @code{pxa3xx} ... instruction register length is 11 bits
@end itemize
@end itemize
@end deffn

To avoid being confused by the variety of ARM based cores, remember
this key point: @emph{ARM is a technology licencing company}.
(See: @url{http://www.arm.com}.)
The CPU name used by OpenOCD will reflect the CPU design that was
licenced, not a vendor brand which incorporates that design.
Name prefixes like arm7, arm9, arm11, and cortex
reflect design generations;
while names like ARMv4, ARMv5, ARMv6, and ARMv7
reflect an architecture version implemented by a CPU design.

@anchor{Target Configuration}
@section Target Configuration

Before creating a ``target'', you must have added its TAP to the scan chain.
When you've added that TAP, you will have a @code{dotted.name}
which is used to set up the CPU support.
The chip-specific configuration file will normally configure its CPU(s)
right after it adds all of the chip's TAPs to the scan chain.

Although you can set up a target in one step, it's often clearer if you
use shorter commands and do it in two steps:  create it, then configure
optional parts.
All operations on the target after it's created will use a new
command, created as part of target creation.

The two main things to configure after target creation are
a work area, which usually has target-specific defaults even
if the board setup code overrides them later;
and event handlers (@pxref{Target Events}), which tend
to be much more board-specific.
The key steps you use might look something like this

@example
target create MyTarget cortex_m3 -chain-position mychip.cpu
$MyTarget configure -work-area-phys 0x08000 -work-area-size 8096
$MyTarget configure -event reset-deassert-pre @{ jtag_rclk 5 @}
$MyTarget configure -event reset-init @{ myboard_reinit @}
@end example

You should specify a working area if you can; typically it uses some
on-chip SRAM.
Such a working area can speed up many things, including bulk
writes to target memory;
flash operations like checking to see if memory needs to be erased;
GDB memory checksumming;
and more.

@quotation Warning
On more complex chips, the work area can become
inaccessible when application code
(such as an operating system)
enables or disables the MMU.
For example, the particular MMU context used to acess the virtual
address will probably matter ... and that context might not have
easy access to other addresses needed.
At this writing, OpenOCD doesn't have much MMU intelligence.
@end quotation

It's often very useful to define a @code{reset-init} event handler.
For systems that are normally used with a boot loader,
common tasks include updating clocks and initializing memory
controllers.
That may be needed to let you write the boot loader into flash,
in order to ``de-brick'' your board; or to load programs into
external DDR memory without having run the boot loader.

@deffn Command {target create} target_name type configparams...
This command creates a GDB debug target that refers to a specific JTAG tap.
It enters that target into a list, and creates a new
command (@command{@var{target_name}}) which is used for various
purposes including additional configuration.

@itemize @bullet
@item @var{target_name} ...  is the name of the debug target.
By convention this should be the same as the @emph{dotted.name}
of the TAP associated with this target, which must be specified here
using the @code{-chain-position @var{dotted.name}} configparam.

This name is also used to create the target object command,
referred to here as @command{$target_name},
and in other places the target needs to be identified.
@item @var{type} ... specifies the target type.  @xref{target types}.
@item @var{configparams} ...  all parameters accepted by
@command{$target_name configure} are permitted.
If the target is big-endian, set it here with @code{-endian big}.
If the variant matters, set it here with @code{-variant}.

You @emph{must} set the @code{-chain-position @var{dotted.name}} here.
@end itemize
@end deffn

@deffn Command {$target_name configure} configparams...
The options accepted by this command may also be
specified as parameters to @command{target create}.
Their values can later be queried one at a time by
using the @command{$target_name cget} command.

@emph{Warning:} changing some of these after setup is dangerous.
For example, moving a target from one TAP to another;
and changing its endianness or variant.

@itemize @bullet

@item @code{-chain-position} @var{dotted.name} -- names the TAP
used to access this target.

@item @code{-endian} (@option{big}|@option{little}) -- specifies
whether the CPU uses big or little endian conventions

@item @code{-event} @var{event_name} @var{event_body} --
@xref{Target Events}.
Note that this updates a list of named event handlers.
Calling this twice with two different event names assigns
two different handlers, but calling it twice with the
same event name assigns only one handler.

@item @code{-variant} @var{name} -- specifies a variant of the target,
which OpenOCD needs to know about.

@item @code{-work-area-backup} (@option{0}|@option{1}) -- says
whether the work area gets backed up; by default,
@emph{it is not backed up.}
When possible, use a working_area that doesn't need to be backed up,
since performing a backup slows down operations.
For example, the beginning of an SRAM block is likely to
be used by most build systems, but the end is often unused.

@item @code{-work-area-size} @var{size} -- specify work are size,
in bytes.  The same size applies regardless of whether its physical
or virtual address is being used.

@item @code{-work-area-phys} @var{address} -- set the work area
base @var{address} to be used when no MMU is active.

@item @code{-work-area-virt} @var{address} -- set the work area
base @var{address} to be used when an MMU is active.
@emph{Do not specify a value for this except on targets with an MMU.}
The value should normally correspond to a static mapping for the
@code{-work-area-phys} address, set up by the current operating system.

@end itemize
@end deffn

@section Other $target_name Commands
@cindex object command

The Tcl/Tk language has the concept of object commands,
and OpenOCD adopts that same model for targets.

A good Tk example is a on screen button.
Once a button is created a button
has a name (a path in Tk terms) and that name is useable as a first
class command. For example in Tk, one can create a button and later
configure it like this:

@example
# Create
button .foobar -background red -command @{ foo @}
# Modify
.foobar configure -foreground blue
# Query
set x [.foobar cget -background]
# Report
puts [format "The button is %s" $x]
@end example

In OpenOCD's terms, the ``target'' is an object just like a Tcl/Tk
button, and its object commands are invoked the same way.

@example
str912.cpu    mww 0x1234 0x42
omap3530.cpu  mww 0x5555 123
@end example

The commands supported by OpenOCD target objects are:

@deffn Command {$target_name arp_examine}
@deffnx Command {$target_name arp_halt}
@deffnx Command {$target_name arp_poll}
@deffnx Command {$target_name arp_reset}
@deffnx Command {$target_name arp_waitstate}
Internal OpenOCD scripts (most notably @file{startup.tcl})
use these to deal with specific reset cases.
They are not otherwise documented here.
@end deffn

@deffn Command {$target_name array2mem} arrayname width address count
@deffnx Command {$target_name mem2array} arrayname width address count
These provide an efficient script-oriented interface to memory.
The @code{array2mem} primitive writes bytes, halfwords, or words;
while @code{mem2array} reads them.
In both cases, the TCL side uses an array, and
the target side uses raw memory.

The efficiency comes from enabling the use of
bulk JTAG data transfer operations.
The script orientation comes from working with data
values that are packaged for use by TCL scripts;
@command{mdw} type primitives only print data they retrieve,
and neither store nor return those values.

@itemize
@item @var{arrayname} ... is the name of an array variable
@item @var{width} ... is 8/16/32 - indicating the memory access size
@item @var{address} ... is the target memory address
@item @var{count} ... is the number of elements to process
@end itemize
@end deffn

@deffn Command {$target_name cget} queryparm
Each configuration parameter accepted by
@command{$target_name configure}
can be individually queried, to return its current value.
The @var{queryparm} is a parameter name
accepted by that command, such as @code{-work-area-phys}.
There are a few special cases:

@itemize @bullet
@item @code{-event} @var{event_name} -- returns the handler for the
event named @var{event_name}.
This is a special case because setting a handler requires
two parameters.
@item @code{-type} -- returns the target type.
This is a special case because this is set using
@command{target create} and can't be changed
using @command{$target_name configure}.
@end itemize

For example, if you wanted to summarize information about
all the targets you might use something like this:

@example
foreach name [target names] @{
    set y [$name cget -endian]
    set z [$name cget -type]
    puts [format "Chip %d is %s, Endian: %s, type: %s" \
                 $x $name $y $z]
@}
@end example
@end deffn

@anchor{target curstate}
@deffn Command {$target_name curstate}
Displays the current target state:
@code{debug-running},
@code{halted},
@code{reset},
@code{running}, or @code{unknown}.
(Also, @pxref{Event Polling}.)
@end deffn

@deffn Command {$target_name eventlist}
Displays a table listing all event handlers
currently associated with this target.
@xref{Target Events}.
@end deffn

@deffn Command {$target_name invoke-event} event_name
Invokes the handler for the event named @var{event_name}.
(This is primarily intended for use by OpenOCD framework
code, for example by the reset code in @file{startup.tcl}.)
@end deffn

@deffn Command {$target_name mdw} addr [count]
@deffnx Command {$target_name mdh} addr [count]
@deffnx Command {$target_name mdb} addr [count]
Display contents of address @var{addr}, as
32-bit words (@command{mdw}), 16-bit halfwords (@command{mdh}),
or 8-bit bytes (@command{mdb}).
If @var{count} is specified, displays that many units.
(If you want to manipulate the data instead of displaying it,
see the @code{mem2array} primitives.)
@end deffn

@deffn Command {$target_name mww} addr word
@deffnx Command {$target_name mwh} addr halfword
@deffnx Command {$target_name mwb} addr byte
Writes the specified @var{word} (32 bits),
@var{halfword} (16 bits), or @var{byte} (8-bit) pattern,
at the specified address @var{addr}.
@end deffn

@anchor{Target Events}
@section Target Events
@cindex target events
@cindex events
At various times, certain things can happen, or you want them to happen.
For example:
@itemize @bullet
@item What should happen when GDB connects? Should your target reset?
@item When GDB tries to flash the target, do you need to enable the flash via a special command?
@item Is using SRST appropriate (and possible) on your system?
Or instead of that, do you need to issue JTAG commands to trigger reset?
SRST usually resets everything on the scan chain, which can be inappropriate.
@item During reset, do you need to write to certain memory locations
to set up system clocks or
to reconfigure the SDRAM?
How about configuring the watchdog timer, or other peripherals,
to stop running while you hold the core stopped for debugging?
@end itemize

All of the above items can be addressed by target event handlers.
These are set up by @command{$target_name configure -event} or
@command{target create ... -event}.

The programmer's model matches the @code{-command} option used in Tcl/Tk
buttons and events.  The two examples below act the same, but one creates
and invokes a small procedure while the other inlines it.

@example
proc my_attach_proc @{ @} @{
    echo "Reset..."
    reset halt
@}
mychip.cpu configure -event gdb-attach my_attach_proc
mychip.cpu configure -event gdb-attach @{
    echo "Reset..."
    reset halt
@}
@end example

The following target events are defined:

@itemize @bullet
@item @b{debug-halted}
@* The target has halted for debug reasons (i.e.: breakpoint)
@item @b{debug-resumed}
@* The target has resumed (i.e.: gdb said run)
@item @b{early-halted}
@* Occurs early in the halt process
@ignore
@item @b{examine-end}
@* Currently not used (goal: when JTAG examine completes)
@item @b{examine-start}
@* Currently not used (goal: when JTAG examine starts)
@end ignore
@item @b{gdb-attach}
@* When GDB connects
@item @b{gdb-detach}
@* When GDB disconnects
@item @b{gdb-end}
@* When the target has halted and GDB is not doing anything (see early halt)
@item @b{gdb-flash-erase-start}
@* Before the GDB flash process tries to erase the flash
@item @b{gdb-flash-erase-end}
@* After the GDB flash process has finished erasing the flash
@item @b{gdb-flash-write-start}
@* Before GDB writes to the flash
@item @b{gdb-flash-write-end}
@* After GDB writes to the flash
@item @b{gdb-start}
@* Before the target steps, gdb is trying to start/resume the target
@item @b{halted}
@* The target has halted
@ignore
@item @b{old-gdb_program_config}
@* DO NOT USE THIS: Used internally
@item @b{old-pre_resume}
@* DO NOT USE THIS: Used internally
@end ignore
@item @b{reset-assert-pre}
@* Issued as part of @command{reset} processing
after @command{reset_init} was triggered
but before either SRST alone is re-asserted on the scan chain,
or @code{reset-assert} is triggered.
@item @b{reset-assert}
@* Issued as part of @command{reset} processing
after @command{reset-assert-pre} was triggered.
When such a handler is present, cores which support this event will use
it instead of asserting SRST.
This support is essential for debugging with JTAG interfaces which
don't include an SRST line (JTAG doesn't require SRST), and for
selective reset on scan chains that have multiple targets.
@item @b{reset-assert-post}
@* Issued as part of @command{reset} processing
after @code{reset-assert} has been triggered.
or the target asserted SRST on the entire scan chain.
@item @b{reset-deassert-pre}
@* Issued as part of @command{reset} processing
after @code{reset-assert-post} has been triggered.
@item @b{reset-deassert-post}
@* Issued as part of @command{reset} processing
after @code{reset-deassert-pre} has been triggered
and (if the target is using it) after SRST has been
released on the scan chain.
@item @b{reset-end}
@* Issued as the final step in @command{reset} processing.
@ignore
@item @b{reset-halt-post}
@* Currently not used
@item @b{reset-halt-pre}
@* Currently not used
@end ignore
@item @b{reset-init}
@* Used by @b{reset init} command for board-specific initialization.
This event fires after @emph{reset-deassert-post}.

This is where you would configure PLLs and clocking, set up DRAM so
you can download programs that don't fit in on-chip SRAM, set up pin
multiplexing, and so on.
(You may be able to switch to a fast JTAG clock rate here, after
the target clocks are fully set up.)
@item @b{reset-start}
@* Issued as part of @command{reset} processing
before @command{reset_init} is called.

This is the most robust place to use @command{jtag_rclk}
or @command{jtag_khz} to switch to a low JTAG clock rate,
when reset disables PLLs needed to use a fast clock.
@ignore
@item @b{reset-wait-pos}
@* Currently not used
@item @b{reset-wait-pre}
@* Currently not used
@end ignore
@item @b{resume-start}
@* Before any target is resumed
@item @b{resume-end}
@* After all targets have resumed
@item @b{resume-ok}
@* Success
@item @b{resumed}
@* Target has resumed
@end itemize


@node Flash Commands
@chapter Flash Commands

OpenOCD has different commands for NOR and NAND flash;
the ``flash'' command works with NOR flash, while
the ``nand'' command works with NAND flash.
This partially reflects different hardware technologies:
NOR flash usually supports direct CPU instruction and data bus access,
while data from a NAND flash must be copied to memory before it can be
used.  (SPI flash must also be copied to memory before use.)
However, the documentation also uses ``flash'' as a generic term;
for example, ``Put flash configuration in board-specific files''.

Flash Steps:
@enumerate
@item Configure via the command @command{flash bank}
@* Do this in a board-specific configuration file,
passing parameters as needed by the driver.
@item Operate on the flash via @command{flash subcommand}
@* Often commands to manipulate the flash are typed by a human, or run
via a script in some automated way.  Common tasks include writing a
boot loader, operating system, or other data.
@item GDB Flashing
@* Flashing via GDB requires the flash be configured via ``flash
bank'', and the GDB flash features be enabled.
@xref{GDB Configuration}.
@end enumerate

Many CPUs have the ablity to ``boot'' from the first flash bank.
This means that misprogramming that bank can ``brick'' a system,
so that it can't boot.
JTAG tools, like OpenOCD, are often then used to ``de-brick'' the
board by (re)installing working boot firmware.

@anchor{NOR Configuration}
@section Flash Configuration Commands
@cindex flash configuration

@deffn {Config Command} {flash bank} name driver base size chip_width bus_width target [driver_options]
Configures a flash bank which provides persistent storage
for addresses from @math{base} to @math{base + size - 1}.
These banks will often be visible to GDB through the target's memory map.
In some cases, configuring a flash bank will activate extra commands;
see the driver-specific documentation.

@itemize @bullet
@item @var{name} ... may be used to reference the flash bank
in other flash commands.
@item @var{driver} ... identifies the controller driver
associated with the flash bank being declared.
This is usually @code{cfi} for external flash, or else
the name of a microcontroller with embedded flash memory.
@xref{Flash Driver List}.
@item @var{base} ... Base address of the flash chip.
@item @var{size} ... Size of the chip, in bytes.
For some drivers, this value is detected from the hardware.
@item @var{chip_width} ... Width of the flash chip, in bytes;
ignored for most microcontroller drivers.
@item @var{bus_width} ... Width of the data bus used to access the
chip, in bytes; ignored for most microcontroller drivers.
@item @var{target} ... Names the target used to issue
commands to the flash controller.
@comment Actually, it's currently a controller-specific parameter...
@item @var{driver_options} ... drivers may support, or require,
additional parameters.  See the driver-specific documentation
for more information.
@end itemize
@quotation Note
This command is not available after OpenOCD initialization has completed.
Use it in board specific configuration files, not interactively.
@end quotation
@end deffn

@comment the REAL name for this command is "ocd_flash_banks"
@comment less confusing would be:  "flash list" (like "nand list")
@deffn Command {flash banks}
Prints a one-line summary of each device that was 
declared using @command{flash bank}, numbered from zero.
Note that this is the @emph{plural} form;
the @emph{singular} form is a very different command.
@end deffn

@deffn Command {flash list}
Retrieves a list of associative arrays for each device that was 
declared using @command{flash bank}, numbered from zero.
This returned list can be manipulated easily from within scripts.
@end deffn

@deffn Command {flash probe} num
Identify the flash, or validate the parameters of the configured flash. Operation
depends on the flash type.
The @var{num} parameter is a value shown by @command{flash banks}.
Most flash commands will implicitly @emph{autoprobe} the bank;
flash drivers can distinguish between probing and autoprobing,
but most don't bother.
@end deffn

@section Erasing, Reading, Writing to Flash
@cindex flash erasing
@cindex flash reading
@cindex flash writing
@cindex flash programming

One feature distinguishing NOR flash from NAND or serial flash technologies
is that for read access, it acts exactly like any other addressible memory.
This means you can use normal memory read commands like @command{mdw} or
@command{dump_image} with it, with no special @command{flash} subcommands.
@xref{Memory access}, and @ref{Image access}.

Write access works differently.  Flash memory normally needs to be erased
before it's written.  Erasing a sector turns all of its bits to ones, and
writing can turn ones into zeroes.  This is why there are special commands
for interactive erasing and writing, and why GDB needs to know which parts
of the address space hold NOR flash memory.

@quotation Note
Most of these erase and write commands leverage the fact that NOR flash
chips consume target address space.  They implicitly refer to the current
JTAG target, and map from an address in that target's address space
back to a flash bank.
@comment In May 2009, those mappings may fail if any bank associated
@comment with that target doesn't succesfuly autoprobe ... bug worth fixing?
A few commands use abstract addressing based on bank and sector numbers,
and don't depend on searching the current target and its address space.
Avoid confusing the two command models.
@end quotation

Some flash chips implement software protection against accidental writes,
since such buggy writes could in some cases ``brick'' a system.
For such systems, erasing and writing may require sector protection to be
disabled first.
Examples include CFI flash such as ``Intel Advanced Bootblock flash'',
and AT91SAM7 on-chip flash.
@xref{flash protect}.

@anchor{flash erase_sector}
@deffn Command {flash erase_sector} num first last
Erase sectors in bank @var{num}, starting at sector @var{first}
up to and including @var{last}.
Sector numbering starts at 0.
Providing a @var{last} sector of @option{last}
specifies "to the end of the flash bank".
The @var{num} parameter is a value shown by @command{flash banks}.
@end deffn

@deffn Command {flash erase_address} address length
Erase sectors starting at @var{address} for @var{length} bytes.
The flash bank to use is inferred from the @var{address}, and
the specified length must stay within that bank.
As a special case, when @var{length} is zero and @var{address} is
the start of the bank, the whole flash is erased.
@end deffn

@deffn Command {flash fillw} address word length
@deffnx Command {flash fillh} address halfword length
@deffnx Command {flash fillb} address byte length
Fills flash memory with the specified @var{word} (32 bits),
@var{halfword} (16 bits), or @var{byte} (8-bit) pattern,
starting at @var{address} and continuing
for @var{length} units (word/halfword/byte).
No erasure is done before writing; when needed, that must be done
before issuing this command.
Writes are done in blocks of up to 1024 bytes, and each write is
verified by reading back the data and comparing it to what was written.
The flash bank to use is inferred from the @var{address} of
each block, and the specified length must stay within that bank.
@end deffn
@comment no current checks for errors if fill blocks touch multiple banks!

@anchor{flash write_bank}
@deffn Command {flash write_bank} num filename offset
Write the binary @file{filename} to flash bank @var{num},
starting at @var{offset} bytes from the beginning of the bank.
The @var{num} parameter is a value shown by @command{flash banks}.
@end deffn

@anchor{flash write_image}
@deffn Command {flash write_image} [erase] [unlock] filename [offset] [type]
Write the image @file{filename} to the current target's flash bank(s).
A relocation @var{offset} may be specified, in which case it is added
to the base address for each section in the image.
The file [@var{type}] can be specified
explicitly as @option{bin} (binary), @option{ihex} (Intel hex),
@option{elf} (ELF file), @option{s19} (Motorola s19).
@option{mem}, or @option{builder}.
The relevant flash sectors will be erased prior to programming
if the @option{erase} parameter is given. If @option{unlock} is
provided, then the flash banks are unlocked before erase and
program. The flash bank to use is inferred from the @var{address} of
each image segment.
@end deffn

@section Other Flash commands
@cindex flash protection

@deffn Command {flash erase_check} num
Check erase state of sectors in flash bank @var{num},
and display that status.
The @var{num} parameter is a value shown by @command{flash banks}.
This is the only operation that
updates the erase state information displayed by @option{flash info}. That means you have
to issue a @command{flash erase_check} command after erasing or programming the device
to get updated information.
(Code execution may have invalidated any state records kept by OpenOCD.)
@end deffn

@deffn Command {flash info} num
Print info about flash bank @var{num}
The @var{num} parameter is a value shown by @command{flash banks}.
The information includes per-sector protect status.
@end deffn

@anchor{flash protect}
@deffn Command {flash protect} num first last (@option{on}|@option{off})
Enable (@option{on}) or disable (@option{off}) protection of flash sectors
in flash bank @var{num}, starting at sector @var{first}
and continuing up to and including @var{last}.
Providing a @var{last} sector of @option{last}
specifies "to the end of the flash bank".
The @var{num} parameter is a value shown by @command{flash banks}.
@end deffn

@deffn Command {flash protect_check} num
Check protection state of sectors in flash bank @var{num}.
The @var{num} parameter is a value shown by @command{flash banks}.
@comment @option{flash erase_sector} using the same syntax.
@end deffn

@anchor{Flash Driver List}
@section Flash Driver List
As noted above, the @command{flash bank} command requires a driver name,
and allows driver-specific options and behaviors.
Some drivers also activate driver-specific commands.

@subsection External Flash

@deffn {Flash Driver} cfi
@cindex Common Flash Interface
@cindex CFI
The ``Common Flash Interface'' (CFI) is the main standard for
external NOR flash chips, each of which connects to a
specific external chip select on the CPU.
Frequently the first such chip is used to boot the system.
Your board's @code{reset-init} handler might need to
configure additional chip selects using other commands (like: @command{mww} to
configure a bus and its timings), or
perhaps configure a GPIO pin that controls the ``write protect'' pin
on the flash chip.
The CFI driver can use a target-specific working area to significantly
speed up operation.

The CFI driver can accept the following optional parameters, in any order:

@itemize
@item @var{jedec_probe} ... is used to detect certain non-CFI flash ROMs,
like AM29LV010 and similar types.
@item @var{x16_as_x8} ... when a 16-bit flash is hooked up to an 8-bit bus.
@end itemize

To configure two adjacent banks of 16 MBytes each, both sixteen bits (two bytes)
wide on a sixteen bit bus:

@example
flash bank cfi 0x00000000 0x01000000 2 2 $_TARGETNAME
flash bank cfi 0x01000000 0x01000000 2 2 $_TARGETNAME
@end example

To configure one bank of 32 MBytes
built from two sixteen bit (two byte) wide parts wired in parallel
to create a thirty-two bit (four byte) bus with doubled throughput:

@example
flash bank cfi 0x00000000 0x02000000 2 4 $_TARGETNAME
@end example

@c "cfi part_id" disabled
@end deffn

@subsection Internal Flash (Microcontrollers)

@deffn {Flash Driver} aduc702x
The ADUC702x analog microcontrollers from Analog Devices
include internal flash and use ARM7TDMI cores.
The aduc702x flash driver works with models ADUC7019 through ADUC7028.
The setup command only requires the @var{target} argument
since all devices in this family have the same memory layout.

@example
flash bank aduc702x 0 0 0 0 $_TARGETNAME
@end example
@end deffn

@deffn {Flash Driver} at91sam3
@cindex at91sam3
All members of the AT91SAM3 microcontroller family from
Atmel include internal flash and use ARM's Cortex-M3 core. The driver
currently (6/22/09) recognizes the AT91SAM3U[1/2/4][C/E] chips. Note
that the driver was orginaly developed and tested using the
AT91SAM3U4E, using a SAM3U-EK eval board. Support for other chips in
the family was cribbed from the data sheet. @emph{Note to future
readers/updaters: Please remove this worrysome comment after other
chips are confirmed.}

The AT91SAM3U4[E/C] (256K) chips have two flash banks; most other chips
have one flash bank.  In all cases the flash banks are at
the following fixed locations:

@example
# Flash bank 0 - all chips
flash bank at91sam3 0x00080000 0 1 1 $_TARGETNAME
# Flash bank 1 - only 256K chips
flash bank at91sam3 0x00100000 0 1 1 $_TARGETNAME
@end example

Internally, the AT91SAM3 flash memory is organized as follows.
Unlike the AT91SAM7 chips, these are not used as parameters
to the @command{flash bank} command:

@itemize
@item @emph{N-Banks:} 256K chips have 2 banks, others have 1 bank.
@item @emph{Bank Size:}  128K/64K Per flash bank
@item @emph{Sectors:} 16 or 8 per bank
@item @emph{SectorSize:} 8K Per Sector
@item @emph{PageSize:} 256 bytes per page. Note that OpenOCD operates on 'sector' sizes, not page sizes.
@end itemize

The AT91SAM3 driver adds some additional commands:

@deffn Command {at91sam3 gpnvm}
@deffnx Command {at91sam3 gpnvm clear} number
@deffnx Command {at91sam3 gpnvm set} number
@deffnx Command {at91sam3 gpnvm show} [@option{all}|number]
With no parameters, @command{show} or @command{show all},
shows the status of all GPNVM bits.
With @command{show} @var{number}, displays that bit.

With @command{set} @var{number} or @command{clear} @var{number},
modifies that GPNVM bit.
@end deffn

@deffn Command {at91sam3 info}
This command attempts to display information about the AT91SAM3
chip. @emph{First} it read the @code{CHIPID_CIDR} [address 0x400e0740, see
Section 28.2.1, page 505 of the AT91SAM3U 29/may/2009 datasheet,
document id: doc6430A] and decodes the values. @emph{Second} it reads the
various clock configuration registers and attempts to display how it
believes the chip is configured. By default, the SLOWCLK is assumed to
be 32768 Hz, see the command @command{at91sam3 slowclk}.
@end deffn

@deffn Command {at91sam3 slowclk} [value]
This command shows/sets the slow clock frequency used in the
@command{at91sam3 info} command calculations above.
@end deffn
@end deffn

@deffn {Flash Driver} at91sam7
All members of the AT91SAM7 microcontroller family from Atmel include
internal flash and use ARM7TDMI cores.  The driver automatically
recognizes a number of these chips using the chip identification
register, and autoconfigures itself.

@example
flash bank at91sam7 0 0 0 0 $_TARGETNAME
@end example

For chips which are not recognized by the controller driver, you must
provide additional parameters in the following order:

@itemize
@item @var{chip_model} ... label used with @command{flash info}
@item @var{banks}
@item @var{sectors_per_bank}
@item @var{pages_per_sector}
@item @var{pages_size}
@item @var{num_nvm_bits}
@item @var{freq_khz} ... required if an external clock is provided,
optional (but recommended) when the oscillator frequency is known
@end itemize

It is recommended that you provide zeroes for all of those values
except the clock frequency, so that everything except that frequency
will be autoconfigured.
Knowing the frequency helps ensure correct timings for flash access.

The flash controller handles erases automatically on a page (128/256 byte)
basis, so explicit erase commands are not necessary for flash programming.
However, there is an ``EraseAll`` command that can erase an entire flash
plane (of up to 256KB), and it will be used automatically when you issue
@command{flash erase_sector} or @command{flash erase_address} commands.

@deffn Command {at91sam7 gpnvm} bitnum (@option{set}|@option{clear})
Set or clear a ``General Purpose Non-Volatle Memory'' (GPNVM)
bit for the processor.   Each processor has a number of such bits,
used for controlling features such as brownout detection (so they
are not truly general purpose).
@quotation Note
This assumes that the first flash bank (number 0) is associated with
the appropriate at91sam7 target.
@end quotation
@end deffn
@end deffn

@deffn {Flash Driver} avr
The AVR 8-bit microcontrollers from Atmel integrate flash memory.
@emph{The current implementation is incomplete.}
@comment - defines mass_erase ... pointless given flash_erase_address
@end deffn

@deffn {Flash Driver} ecosflash
@emph{No idea what this is...}
The @var{ecosflash} driver defines one mandatory parameter,
the name of a modules of target code which is downloaded
and executed.
@end deffn

@deffn {Flash Driver} lpc2000
Most members of the LPC1700 and LPC2000 microcontroller families from NXP
include internal flash and use Cortex-M3 (LPC1700) or ARM7TDMI (LPC2000) cores.

@quotation Note
There are LPC2000 devices which are not supported by the @var{lpc2000}
driver:
The LPC2888 is supported by the @var{lpc288x} driver.
The LPC29xx family is supported by the @var{lpc2900} driver.
@end quotation

The @var{lpc2000} driver defines two mandatory and one optional parameters,
which must appear in the following order:

@itemize
@item @var{variant} ... required, may be
@var{lpc2000_v1} (older LPC21xx and LPC22xx)
@var{lpc2000_v2} (LPC213x, LPC214x, LPC210[123], LPC23xx and LPC24xx)
or @var{lpc1700} (LPC175x and LPC176x)
@item @var{clock_kHz} ... the frequency, in kiloHertz,
at which the core is running
@item @var{calc_checksum} ... optional (but you probably want to provide this!),
telling the driver to calculate a valid checksum for the exception vector table.
@end itemize

LPC flashes don't require the chip and bus width to be specified.

@example
flash bank lpc2000 0x0 0x7d000 0 0 $_TARGETNAME \
      lpc2000_v2 14765 calc_checksum
@end example

@deffn {Command} {lpc2000 part_id} bank
Displays the four byte part identifier associated with
the specified flash @var{bank}.
@end deffn
@end deffn

@deffn {Flash Driver} lpc288x
The LPC2888 microcontroller from NXP needs slightly different flash
support from its lpc2000 siblings.
The @var{lpc288x} driver defines one mandatory parameter,
the programming clock rate in Hz.
LPC flashes don't require the chip and bus width to be specified.

@example
flash bank lpc288x 0 0 0 0 $_TARGETNAME 12000000
@end example
@end deffn

@deffn {Flash Driver} lpc2900
This driver supports the LPC29xx ARM968E based microcontroller family
from NXP.

The predefined parameters @var{base}, @var{size}, @var{chip_width} and
@var{bus_width} of the @code{flash bank} command are ignored. Flash size and
sector layout are auto-configured by the driver.
The driver has one additional mandatory parameter: The CPU clock rate
(in kHz) at the time the flash operations will take place. Most of the time this
will not be the crystal frequency, but a higher PLL frequency. The
@code{reset-init} event handler in the board script is usually the place where
you start the PLL.

The driver rejects flashless devices (currently the LPC2930).

The EEPROM in LPC2900 devices is not mapped directly into the address space.
It must be handled much more like NAND flash memory, and will therefore be
handled by a separate @code{lpc2900_eeprom} driver (not yet available).

Sector protection in terms of the LPC2900 is handled transparently. Every time a
sector needs to be erased or programmed, it is automatically unprotected.
What is shown as protection status in the @code{flash info} command, is
actually the LPC2900 @emph{sector security}. This is a mechanism to prevent a
sector from ever being erased or programmed again. As this is an irreversible
mechanism, it is handled by a special command (@code{lpc2900 secure_sector}),
and not by the standard @code{flash protect} command.

Example for a 125 MHz clock frequency:
@example
flash bank lpc2900 0 0 0 0 $_TARGETNAME 125000
@end example

Some @code{lpc2900}-specific commands are defined. In the following command list,
the @var{bank} parameter is the bank number as obtained by the
@code{flash banks} command.

@deffn Command {lpc2900 signature} bank
Calculates a 128-bit hash value, the @emph{signature}, from the whole flash
content. This is a hardware feature of the flash block, hence the calculation is
very fast. You may use this to verify the content of a programmed device against
a known signature.
Example:
@example
lpc2900 signature 0
  signature: 0x5f40cdc8:0xc64e592e:0x10490f89:0x32a0f317
@end example
@end deffn

@deffn Command {lpc2900 read_custom} bank filename
Reads the 912 bytes of customer information from the flash index sector, and
saves it to a file in binary format.
Example:
@example
lpc2900 read_custom 0 /path_to/customer_info.bin
@end example
@end deffn

The index sector of the flash is a @emph{write-only} sector. It cannot be
erased! In order to guard against unintentional write access, all following
commands need to be preceeded by a successful call to the @code{password}
command:

@deffn Command {lpc2900 password} bank password
You need to use this command right before each of the following commands:
@code{lpc2900 write_custom}, @code{lpc2900 secure_sector},
@code{lpc2900 secure_jtag}.

The password string is fixed to "I_know_what_I_am_doing".
Example:
@example
lpc2900 password 0 I_know_what_I_am_doing
  Potentially dangerous operation allowed in next command!
@end example
@end deffn

@deffn Command {lpc2900 write_custom} bank filename type
Writes the content of the file into the customer info space of the flash index
sector. The filetype can be specified with the @var{type} field. Possible values
for @var{type} are: @var{bin} (binary), @var{ihex} (Intel hex format),
@var{elf} (ELF binary) or @var{s19} (Motorola S-records). The file must
contain a single section, and the contained data length must be exactly
912 bytes.
@quotation Attention
This cannot be reverted! Be careful!
@end quotation
Example:
@example
lpc2900 write_custom 0 /path_to/customer_info.bin bin
@end example
@end deffn

@deffn Command {lpc2900 secure_sector} bank first last
Secures the sector range from @var{first} to @var{last} (including) against
further program and erase operations. The sector security will be effective
after the next power cycle.
@quotation Attention
This cannot be reverted! Be careful!
@end quotation
Secured sectors appear as @emph{protected} in the @code{flash info} command.
Example:
@example
lpc2900 secure_sector 0 1 1
flash info 0
  #0 : lpc2900 at 0x20000000, size 0x000c0000, (...)
          #  0: 0x00000000 (0x2000 8kB) not protected
          #  1: 0x00002000 (0x2000 8kB) protected
          #  2: 0x00004000 (0x2000 8kB) not protected
@end example
@end deffn

@deffn Command {lpc2900 secure_jtag} bank
Irreversibly disable the JTAG port. The new JTAG security setting will be
effective after the next power cycle.
@quotation Attention
This cannot be reverted! Be careful!
@end quotation
Examples:
@example
lpc2900 secure_jtag 0
@end example
@end deffn
@end deffn

@deffn {Flash Driver} ocl
@emph{No idea what this is, other than using some arm7/arm9 core.}

@example
flash bank ocl 0 0 0 0 $_TARGETNAME
@end example
@end deffn

@deffn {Flash Driver} pic32mx
The PIC32MX microcontrollers are based on the MIPS 4K cores,
and integrate flash memory.
@emph{The current implementation is incomplete.}

@example
flash bank pix32mx 0 0 0 0 $_TARGETNAME
@end example

@comment numerous *disabled* commands are defined:
@comment - chip_erase ... pointless given flash_erase_address
@comment - lock, unlock ... pointless given protect on/off (yes?)
@comment - pgm_word ... shouldn't bank be deduced from address??
Some pic32mx-specific commands are defined:
@deffn Command {pic32mx pgm_word} address value bank
Programs the specified 32-bit @var{value} at the given @var{address}
in the specified chip @var{bank}.
@end deffn
@end deffn

@deffn {Flash Driver} stellaris
All members of the Stellaris LM3Sxxx microcontroller family from
Texas Instruments
include internal flash and use ARM Cortex M3 cores.
The driver automatically recognizes a number of these chips using
the chip identification register, and autoconfigures itself.
@footnote{Currently there is a @command{stellaris mass_erase} command.
That seems pointless since the same effect can be had using the
standard @command{flash erase_address} command.}

@example
flash bank stellaris 0 0 0 0 $_TARGETNAME
@end example
@end deffn

@deffn {Flash Driver} stm32x
All members of the STM32 microcontroller family from ST Microelectronics
include internal flash and use ARM Cortex M3 cores.
The driver automatically recognizes a number of these chips using
the chip identification register, and autoconfigures itself.

@example
flash bank stm32x 0 0 0 0 $_TARGETNAME
@end example

Some stm32x-specific commands
@footnote{Currently there is a @command{stm32x mass_erase} command.
That seems pointless since the same effect can be had using the
standard @command{flash erase_address} command.}
are defined:

@deffn Command {stm32x lock} num
Locks the entire stm32 device.
The @var{num} parameter is a value shown by @command{flash banks}.
@end deffn

@deffn Command {stm32x unlock} num
Unlocks the entire stm32 device.
The @var{num} parameter is a value shown by @command{flash banks}.
@end deffn

@deffn Command {stm32x options_read} num
Read and display the stm32 option bytes written by
the @command{stm32x options_write} command.
The @var{num} parameter is a value shown by @command{flash banks}.
@end deffn

@deffn Command {stm32x options_write} num (@option{SWWDG}|@option{HWWDG}) (@option{RSTSTNDBY}|@option{NORSTSTNDBY}) (@option{RSTSTOP}|@option{NORSTSTOP})
Writes the stm32 option byte with the specified values.
The @var{num} parameter is a value shown by @command{flash banks}.
@end deffn
@end deffn

@deffn {Flash Driver} str7x
All members of the STR7 microcontroller family from ST Microelectronics
include internal flash and use ARM7TDMI cores.
The @var{str7x} driver defines one mandatory parameter, @var{variant},
which is either @code{STR71x}, @code{STR73x} or @code{STR75x}.

@example
flash bank str7x 0x40000000 0x00040000 0 0 $_TARGETNAME STR71x
@end example

@deffn Command {str7x disable_jtag} bank
Activate the Debug/Readout protection mechanism
for the specified flash bank.
@end deffn
@end deffn

@deffn {Flash Driver} str9x
Most members of the STR9 microcontroller family from ST Microelectronics
include internal flash and use ARM966E cores.
The str9 needs the flash controller to be configured using
the @command{str9x flash_config} command prior to Flash programming.

@example
flash bank str9x 0x40000000 0x00040000 0 0 $_TARGETNAME
str9x flash_config 0 4 2 0 0x80000
@end example

@deffn Command {str9x flash_config} num bbsr nbbsr bbadr nbbadr
Configures the str9 flash controller.
The @var{num} parameter is a value shown by @command{flash banks}.

@itemize @bullet
@item @var{bbsr} - Boot Bank Size register
@item @var{nbbsr} - Non Boot Bank Size register
@item @var{bbadr} - Boot Bank Start Address register
@item @var{nbbadr} - Boot Bank Start Address register
@end itemize
@end deffn

@end deffn

@deffn {Flash Driver} tms470
Most members of the TMS470 microcontroller family from Texas Instruments
include internal flash and use ARM7TDMI cores.
This driver doesn't require the chip and bus width to be specified.

Some tms470-specific commands are defined:

@deffn Command {tms470 flash_keyset} key0 key1 key2 key3
Saves programming keys in a register, to enable flash erase and write commands.
@end deffn

@deffn Command {tms470 osc_mhz} clock_mhz
Reports the clock speed, which is used to calculate timings.
@end deffn

@deffn Command {tms470 plldis} (0|1)
Disables (@var{1}) or enables (@var{0}) use of the PLL to speed up
the flash clock.
@end deffn
@end deffn

@subsection str9xpec driver
@cindex str9xpec

Here is some background info to help
you better understand how this driver works. OpenOCD has two flash drivers for
the str9:
@enumerate
@item
Standard driver @option{str9x} programmed via the str9 core. Normally used for
flash programming as it is faster than the @option{str9xpec} driver.
@item
Direct programming @option{str9xpec} using the flash controller. This is an
ISC compilant (IEEE 1532) tap connected in series with the str9 core. The str9
core does not need to be running to program using this flash driver. Typical use
for this driver is locking/unlocking the target and programming the option bytes.
@end enumerate

Before we run any commands using the @option{str9xpec} driver we must first disable
the str9 core. This example assumes the @option{str9xpec} driver has been
configured for flash bank 0.
@example
# assert srst, we do not want core running
# while accessing str9xpec flash driver
jtag_reset 0 1
# turn off target polling
poll off
# disable str9 core
str9xpec enable_turbo 0
# read option bytes
str9xpec options_read 0
# re-enable str9 core
str9xpec disable_turbo 0
poll on
reset halt
@end example
The above example will read the str9 option bytes.
When performing a unlock remember that you will not be able to halt the str9 - it
has been locked. Halting the core is not required for the @option{str9xpec} driver
as mentioned above, just issue the commands above manually or from a telnet prompt.

@deffn {Flash Driver} str9xpec
Only use this driver for locking/unlocking the device or configuring the option bytes.
Use the standard str9 driver for programming.
Before using the flash commands the turbo mode must be enabled using the
@command{str9xpec enable_turbo} command.

Several str9xpec-specific commands are defined:

@deffn Command {str9xpec disable_turbo} num
Restore the str9 into JTAG chain.
@end deffn

@deffn Command {str9xpec enable_turbo} num
Enable turbo mode, will simply remove the str9 from the chain and talk
directly to the embedded flash controller.
@end deffn

@deffn Command {str9xpec lock} num
Lock str9 device. The str9 will only respond to an unlock command that will
erase the device.
@end deffn

@deffn Command {str9xpec part_id} num
Prints the part identifier for bank @var{num}.
@end deffn

@deffn Command {str9xpec options_cmap} num (@option{bank0}|@option{bank1})
Configure str9 boot bank.
@end deffn

@deffn Command {str9xpec options_lvdsel} num (@option{vdd}|@option{vdd_vddq})
Configure str9 lvd source.
@end deffn

@deffn Command {str9xpec options_lvdthd} num (@option{2.4v}|@option{2.7v})
Configure str9 lvd threshold.
@end deffn

@deffn Command {str9xpec options_lvdwarn} bank (@option{vdd}|@option{vdd_vddq})
Configure str9 lvd reset warning source.
@end deffn

@deffn Command {str9xpec options_read} num
Read str9 option bytes.
@end deffn

@deffn Command {str9xpec options_write} num
Write str9 option bytes.
@end deffn

@deffn Command {str9xpec unlock} num
unlock str9 device.
@end deffn

@end deffn


@section mFlash

@subsection mFlash Configuration
@cindex mFlash Configuration

@deffn {Config Command} {mflash bank} soc base RST_pin target
Configures a mflash for @var{soc} host bank at
address @var{base}.
The pin number format depends on the host GPIO naming convention.
Currently, the mflash driver supports s3c2440 and pxa270.

Example for s3c2440 mflash where @var{RST pin} is GPIO B1:

@example
mflash bank s3c2440 0x10000000 1b 0
@end example

Example for pxa270 mflash where @var{RST pin} is GPIO 43:

@example
mflash bank pxa270 0x08000000 43 0
@end example
@end deffn

@subsection mFlash commands
@cindex mFlash commands

@deffn Command {mflash config pll} frequency
Configure mflash PLL.
The @var{frequency} is the mflash input frequency, in Hz.
Issuing this command will erase mflash's whole internal nand and write new pll.
After this command, mflash needs power-on-reset for normal operation.
If pll was newly configured, storage and boot(optional) info also need to be update.
@end deffn

@deffn Command {mflash config boot}
Configure bootable option.
If bootable option is set, mflash offer the first 8 sectors
(4kB) for boot.
@end deffn

@deffn Command {mflash config storage}
Configure storage information.
For the normal storage operation, this information must be
written.
@end deffn

@deffn Command {mflash dump} num filename offset size
Dump @var{size} bytes, starting at @var{offset} bytes from the
beginning of the bank @var{num}, to the file named @var{filename}.
@end deffn

@deffn Command {mflash probe}
Probe mflash.
@end deffn

@deffn Command {mflash write} num filename offset
Write the binary file @var{filename} to mflash bank @var{num}, starting at
@var{offset} bytes from the beginning of the bank.
@end deffn

@node NAND Flash Commands
@chapter NAND Flash Commands
@cindex NAND

Compared to NOR or SPI flash, NAND devices are inexpensive
and high density.  Today's NAND chips, and multi-chip modules,
commonly hold multiple GigaBytes of data.

NAND chips consist of a number of ``erase blocks'' of a given
size (such as 128 KBytes), each of which is divided into a
number of pages (of perhaps 512 or 2048 bytes each).  Each
page of a NAND flash has an ``out of band'' (OOB) area to hold
Error Correcting Code (ECC) and other metadata, usually 16 bytes
of OOB for every 512 bytes of page data.

One key characteristic of NAND flash is that its error rate
is higher than that of NOR flash.  In normal operation, that
ECC is used to correct and detect errors.  However, NAND
blocks can also wear out and become unusable; those blocks
are then marked "bad".  NAND chips are even shipped from the
manufacturer with a few bad blocks.  The highest density chips
use a technology (MLC) that wears out more quickly, so ECC
support is increasingly important as a way to detect blocks
that have begun to fail, and help to preserve data integrity
with techniques such as wear leveling.

Software is used to manage the ECC.  Some controllers don't
support ECC directly; in those cases, software ECC is used.
Other controllers speed up the ECC calculations with hardware.
Single-bit error correction hardware is routine.  Controllers
geared for newer MLC chips may correct 4 or more errors for
every 512 bytes of data.

You will need to make sure that any data you write using
OpenOCD includes the apppropriate kind of ECC.  For example,
that may mean passing the @code{oob_softecc} flag when
writing NAND data, or ensuring that the correct hardware
ECC mode is used.

The basic steps for using NAND devices include:
@enumerate
@item Declare via the command @command{nand device}
@* Do this in a board-specific configuration file,
passing parameters as needed by the controller.
@item Configure each device using @command{nand probe}.
@* Do this only after the associated target is set up,
such as in its reset-init script or in procures defined
to access that device.
@item Operate on the flash via @command{nand subcommand}
@* Often commands to manipulate the flash are typed by a human, or run
via a script in some automated way.  Common task include writing a
boot loader, operating system, or other data needed to initialize or
de-brick a board.
@end enumerate

@b{NOTE:} At the time this text was written, the largest NAND
flash fully supported by OpenOCD is 2 GiBytes (16 GiBits).
This is because the variables used to hold offsets and lengths
are only 32 bits wide.
(Larger chips may work in some cases, unless an offset or length
is larger than 0xffffffff, the largest 32-bit unsigned integer.)
Some larger devices will work, since they are actually multi-chip
modules with two smaller chips and individual chipselect lines.

@anchor{NAND Configuration}
@section NAND Configuration Commands
@cindex NAND configuration

NAND chips must be declared in configuration scripts,
plus some additional configuration that's done after
OpenOCD has initialized.

@deffn {Config Command} {nand device} name controller target [configparams...]
Declares a NAND device, which can be read and written to
after it has been configured through @command{nand probe}.
In OpenOCD, devices are single chips; this is unlike some
operating systems, which may manage multiple chips as if
they were a single (larger) device.
In some cases, configuring a device will activate extra
commands; see the controller-specific documentation.

@b{NOTE:} This command is not available after OpenOCD
initialization has completed.  Use it in board specific
configuration files, not interactively.

@itemize @bullet
@item @var{name} ... may be used to reference the NAND bank
in other commands.
@item @var{controller} ... identifies the controller driver
associated with the NAND device being declared.
@xref{NAND Driver List}.
@item @var{target} ... names the target used when issuing
commands to the NAND controller.
@comment Actually, it's currently a controller-specific parameter...
@item @var{configparams} ... controllers may support, or require,
additional parameters.  See the controller-specific documentation
for more information.
@end itemize
@end deffn

@deffn Command {nand list}
Prints a summary of each device declared
using @command{nand device}, numbered from zero.
Note that un-probed devices show no details.
@example
> nand list
#0: NAND 1GiB 3,3V 8-bit (Micron) pagesize: 2048, buswidth: 8,
        blocksize: 131072, blocks: 8192
#1: NAND 1GiB 3,3V 8-bit (Micron) pagesize: 2048, buswidth: 8,
        blocksize: 131072, blocks: 8192
>
@end example
@end deffn

@deffn Command {nand probe} num
Probes the specified device to determine key characteristics
like its page and block sizes, and how many blocks it has.
The @var{num} parameter is the value shown by @command{nand list}.
You must (successfully) probe a device before you can use
it with most other NAND commands.
@end deffn

@section Erasing, Reading, Writing to NAND Flash

@deffn Command {nand dump} num filename offset length [oob_option]
@cindex NAND reading
Reads binary data from the NAND device and writes it to the file,
starting at the specified offset.
The @var{num} parameter is the value shown by @command{nand list}.

Use a complete path name for @var{filename}, so you don't depend
on the directory used to start the OpenOCD server.

The @var{offset} and @var{length} must be exact multiples of the
device's page size.  They describe a data region; the OOB data
associated with each such page may also be accessed.

@b{NOTE:} At the time this text was written, no error correction
was done on the data that's read, unless raw access was disabled
and the underlying NAND controller driver had a @code{read_page}
method which handled that error correction.

By default, only page data is saved to the specified file.
Use an @var{oob_option} parameter to save OOB data:
@itemize @bullet
@item no oob_* parameter
@*Output file holds only page data; OOB is discarded.
@item @code{oob_raw}
@*Output file interleaves page data and OOB data;
the file will be longer than "length" by the size of the
spare areas associated with each data page.
Note that this kind of "raw" access is different from
what's implied by @command{nand raw_access}, which just
controls whether a hardware-aware access method is used.
@item @code{oob_only}
@*Output file has only raw OOB data, and will
be smaller than "length" since it will contain only the
spare areas associated with each data page.
@end itemize
@end deffn

@deffn Command {nand erase} num [offset length]
@cindex NAND erasing
@cindex NAND programming
Erases blocks on the specified NAND device, starting at the
specified @var{offset} and continuing for @var{length} bytes.
Both of those values must be exact multiples of the device's
block size, and the region they specify must fit entirely in the chip.
If those parameters are not specified,
the whole NAND chip will be erased.
The @var{num} parameter is the value shown by @command{nand list}.

@b{NOTE:} This command will try to erase bad blocks, when told
to do so, which will probably invalidate the manufacturer's bad
block marker.
For the remainder of the current server session, @command{nand info}
will still report that the block ``is'' bad.
@end deffn

@deffn Command {nand write} num filename offset [option...]
@cindex NAND writing
@cindex NAND programming
Writes binary data from the file into the specified NAND device,
starting at the specified offset.  Those pages should already
have been erased; you can't change zero bits to one bits.
The @var{num} parameter is the value shown by @command{nand list}.

Use a complete path name for @var{filename}, so you don't depend
on the directory used to start the OpenOCD server.

The @var{offset} must be an exact multiple of the device's page size.
All data in the file will be written, assuming it doesn't run
past the end of the device.
Only full pages are written, and any extra space in the last
page will be filled with 0xff bytes.  (That includes OOB data,
if that's being written.)

@b{NOTE:} At the time this text was written, bad blocks are
ignored.  That is, this routine will not skip bad blocks,
but will instead try to write them.  This can cause problems.

Provide at most one @var{option} parameter.  With some
NAND drivers, the meanings of these parameters may change
if @command{nand raw_access} was used to disable hardware ECC.
@itemize @bullet
@item no oob_* parameter
@*File has only page data, which is written.
If raw acccess is in use, the OOB area will not be written.
Otherwise, if the underlying NAND controller driver has
a @code{write_page} routine, that routine may write the OOB
with hardware-computed ECC data.
@item @code{oob_only}
@*File has only raw OOB data, which is written to the OOB area.
Each page's data area stays untouched.  @i{This can be a dangerous
option}, since it can invalidate the ECC data.
You may need to force raw access to use this mode.
@item @code{oob_raw}
@*File interleaves data and OOB data, both of which are written
If raw access is enabled, the data is written first, then the
un-altered OOB.
Otherwise, if the underlying NAND controller driver has
a @code{write_page} routine, that routine may modify the OOB
before it's written, to include hardware-computed ECC data.
@item @code{oob_softecc}
@*File has only page data, which is written.
The OOB area is filled with 0xff, except for a standard 1-bit
software ECC code stored in conventional locations.
You might need to force raw access to use this mode, to prevent
the underlying driver from applying hardware ECC.
@item @code{oob_softecc_kw}
@*File has only page data, which is written.
The OOB area is filled with 0xff, except for a 4-bit software ECC
specific to the boot ROM in Marvell Kirkwood SoCs.
You might need to force raw access to use this mode, to prevent
the underlying driver from applying hardware ECC.
@end itemize
@end deffn

@deffn Command {nand verify} num filename offset [option...]
@cindex NAND verification
@cindex NAND programming
Verify the binary data in the file has been programmed to the
specified NAND device, starting at the specified offset.
The @var{num} parameter is the value shown by @command{nand list}.

Use a complete path name for @var{filename}, so you don't depend
on the directory used to start the OpenOCD server.

The @var{offset} must be an exact multiple of the device's page size.
All data in the file will be read and compared to the contents of the
flash, assuming it doesn't run past the end of the device.
As with @command{nand write}, only full pages are verified, so any extra
space in the last page will be filled with 0xff bytes.

The same @var{options} accepted by @command{nand write},
and the file will be processed similarly to produce the buffers that
can be compared against the contents produced from @command{nand dump}.

@b{NOTE:} This will not work when the underlying NAND controller
driver's @code{write_page} routine must update the OOB with a
hardward-computed ECC before the data is written.  This limitation may
be removed in a future release.
@end deffn

@section Other NAND commands
@cindex NAND other commands

@deffn Command {nand check_bad_blocks} [offset length]
Checks for manufacturer bad block markers on the specified NAND
device.  If no parameters are provided, checks the whole
device; otherwise, starts at the specified @var{offset} and
continues for @var{length} bytes.
Both of those values must be exact multiples of the device's
block size, and the region they specify must fit entirely in the chip.
The @var{num} parameter is the value shown by @command{nand list}.

@b{NOTE:} Before using this command you should force raw access
with @command{nand raw_access enable} to ensure that the underlying
driver will not try to apply hardware ECC.
@end deffn

@deffn Command {nand info} num
The @var{num} parameter is the value shown by @command{nand list}.
This prints the one-line summary from "nand list", plus for
devices which have been probed this also prints any known
status for each block.
@end deffn

@deffn Command {nand raw_access} num (@option{enable}|@option{disable})
Sets or clears an flag affecting how page I/O is done.
The @var{num} parameter is the value shown by @command{nand list}.

This flag is cleared (disabled) by default, but changing that
value won't affect all NAND devices.  The key factor is whether
the underlying driver provides @code{read_page} or @code{write_page}
methods.  If it doesn't provide those methods, the setting of
this flag is irrelevant; all access is effectively ``raw''.

When those methods exist, they are normally used when reading
data (@command{nand dump} or reading bad block markers) or
writing it (@command{nand write}).  However, enabling
raw access (setting the flag) prevents use of those methods,
bypassing hardware ECC logic.
@i{This can be a dangerous option}, since writing blocks
with the wrong ECC data can cause them to be marked as bad.
@end deffn

@anchor{NAND Driver List}
@section NAND Driver List
As noted above, the @command{nand device} command allows
driver-specific options and behaviors.
Some controllers also activate controller-specific commands.

@deffn {NAND Driver} davinci
This driver handles the NAND controllers found on DaVinci family
chips from Texas Instruments.
It takes three extra parameters:
address of the NAND chip;
hardware ECC mode to use (@option{hwecc1},
@option{hwecc4}, @option{hwecc4_infix});
address of the AEMIF controller on this processor.
@example
nand device davinci dm355.arm 0x02000000 hwecc4 0x01e10000
@end example
All DaVinci processors support the single-bit ECC hardware,
and newer ones also support the four-bit ECC hardware.
The @code{write_page} and @code{read_page} methods are used
to implement those ECC modes, unless they are disabled using
the @command{nand raw_access} command.
@end deffn

@deffn {NAND Driver} lpc3180
These controllers require an extra @command{nand device}
parameter:  the clock rate used by the controller.
@deffn Command {lpc3180 select} num [mlc|slc]
Configures use of the MLC or SLC controller mode.
MLC implies use of hardware ECC.
The @var{num} parameter is the value shown by @command{nand list}.
@end deffn

At this writing, this driver includes @code{write_page}
and @code{read_page} methods.  Using @command{nand raw_access}
to disable those methods will prevent use of hardware ECC
in the MLC controller mode, but won't change SLC behavior.
@end deffn
@comment current lpc3180 code won't issue 5-byte address cycles

@deffn {NAND Driver} orion
These controllers require an extra @command{nand device}
parameter:  the address of the controller.
@example
nand device orion 0xd8000000
@end example
These controllers don't define any specialized commands.
At this writing, their drivers don't include @code{write_page}
or @code{read_page} methods, so @command{nand raw_access} won't
change any behavior.
@end deffn

@deffn {NAND Driver} s3c2410
@deffnx {NAND Driver} s3c2412
@deffnx {NAND Driver} s3c2440
@deffnx {NAND Driver} s3c2443
These S3C24xx family controllers don't have any special
@command{nand device} options, and don't define any
specialized commands.
At this writing, their drivers don't include @code{write_page}
or @code{read_page} methods, so @command{nand raw_access} won't
change any behavior.
@end deffn

@node PLD/FPGA Commands
@chapter PLD/FPGA Commands
@cindex PLD
@cindex FPGA

Programmable Logic Devices (PLDs) and the more flexible
Field Programmable Gate Arrays (FPGAs) are both types of programmable hardware.
OpenOCD can support programming them.
Although PLDs are generally restrictive (cells are less functional, and
there are no special purpose cells for memory or computational tasks),
they share the same OpenOCD infrastructure.
Accordingly, both are called PLDs here.

@section PLD/FPGA Configuration and Commands

As it does for JTAG TAPs, debug targets, and flash chips (both NOR and NAND),
OpenOCD maintains a list of PLDs available for use in various commands.
Also, each such PLD requires a driver.

They are referenced by the number shown by the @command{pld devices} command,
and new PLDs are defined by @command{pld device driver_name}.

@deffn {Config Command} {pld device} driver_name tap_name [driver_options]
Defines a new PLD device, supported by driver @var{driver_name},
using the TAP named @var{tap_name}.
The driver may make use of any @var{driver_options} to configure its
behavior.
@end deffn

@deffn {Command} {pld devices}
Lists the PLDs and their numbers.
@end deffn

@deffn {Command} {pld load} num filename
Loads the file @file{filename} into the PLD identified by @var{num}.
The file format must be inferred by the driver.
@end deffn

@section PLD/FPGA Drivers, Options, and Commands

Drivers may support PLD-specific options to the @command{pld device}
definition command, and may also define commands usable only with
that particular type of PLD.

@deffn {FPGA Driver} virtex2
Virtex-II is a family of FPGAs sold by Xilinx.
It supports the IEEE 1532 standard for In-System Configuration (ISC).
No driver-specific PLD definition options are used,
and one driver-specific command is defined.

@deffn {Command} {virtex2 read_stat} num
Reads and displays the Virtex-II status register (STAT)
for FPGA @var{num}.
@end deffn
@end deffn

@node General Commands
@chapter General Commands
@cindex commands

The commands documented in this chapter here are common commands that
you, as a human, may want to type and see the output of. Configuration type
commands are documented elsewhere.

Intent:
@itemize @bullet
@item @b{Source Of Commands}
@* OpenOCD commands can occur in a configuration script (discussed
elsewhere) or typed manually by a human or supplied programatically,
or via one of several TCP/IP Ports.

@item @b{From the human}
@* A human should interact with the telnet interface (default port: 4444)
or via GDB (default port 3333).

To issue commands from within a GDB session, use the @option{monitor}
command, e.g. use @option{monitor poll} to issue the @option{poll}
command. All output is relayed through the GDB session.

@item @b{Machine Interface}
The Tcl interface's intent is to be a machine interface. The default Tcl
port is 5555.
@end itemize


@section Daemon Commands

@deffn {Command} exit
Exits the current telnet session.
@end deffn

@c note EXTREMELY ANNOYING word wrap at column 75
@c even when lines are e.g. 100+ columns ...
@c coded in startup.tcl
@deffn {Command} help [string]
With no parameters, prints help text for all commands.
Otherwise, prints each helptext containing @var{string}.
Not every command provides helptext.
@end deffn

@deffn Command sleep msec [@option{busy}]
Wait for at least @var{msec} milliseconds before resuming.
If @option{busy} is passed, busy-wait instead of sleeping.
(This option is strongly discouraged.)
Useful in connection with script files
(@command{script} command and @command{target_name} configuration).
@end deffn

@deffn Command shutdown
Close the OpenOCD daemon, disconnecting all clients (GDB, telnet, other).
@end deffn

@anchor{debug_level}
@deffn Command debug_level [n]
@cindex message level
Display debug level.
If @var{n} (from 0..3) is provided, then set it to that level.
This affects the kind of messages sent to the server log.
Level 0 is error messages only;
level 1 adds warnings;
level 2 adds informational messages;
and level 3 adds debugging messages.
The default is level 2, but that can be overridden on
the command line along with the location of that log
file (which is normally the server's standard output).
@xref{Running}.
@end deffn

@deffn Command fast (@option{enable}|@option{disable})
Default disabled.
Set default behaviour of OpenOCD to be "fast and dangerous".

At this writing, this only affects the defaults for two ARM7/ARM9 parameters:
fast memory access, and DCC downloads.  Those parameters may still be
individually overridden.

The target specific "dangerous" optimisation tweaking options may come and go
as more robust and user friendly ways are found to ensure maximum throughput
and robustness with a minimum of configuration.

Typically the "fast enable" is specified first on the command line:

@example
openocd -c "fast enable" -c "interface dummy" -f target/str710.cfg
@end example
@end deffn

@deffn Command echo message
Logs a message at "user" priority.
Output @var{message} to stdout.
@example
echo "Downloading kernel -- please wait"
@end example
@end deffn

@deffn Command log_output [filename]
Redirect logging to @var{filename};
the initial log output channel is stderr.
@end deffn

@anchor{Target State handling}
@section Target State handling
@cindex reset
@cindex halt
@cindex target initialization

In this section ``target'' refers to a CPU configured as
shown earlier (@pxref{CPU Configuration}).
These commands, like many, implicitly refer to
a current target which is used to perform the
various operations.  The current target may be changed
by using @command{targets} command with the name of the
target which should become current.

@deffn Command reg [(number|name) [value]]
Access a single register by @var{number} or by its @var{name}.
The target must generally be halted before access to CPU core
registers is allowed.  Depending on the hardware, some other
registers may be accessible while the target is running.

@emph{With no arguments}:
list all available registers for the current target,
showing number, name, size, value, and cache status.
For valid entries, a value is shown; valid entries
which are also dirty (and will be written back later)
are flagged as such.

@emph{With number/name}: display that register's value.

@emph{With both number/name and value}: set register's value.
Writes may be held in a writeback cache internal to OpenOCD,
so that setting the value marks the register as dirty instead
of immediately flushing that value.  Resuming CPU execution
(including by single stepping) or otherwise activating the
relevant module will flush such values.

Cores may have surprisingly many registers in their
Debug and trace infrastructure:

@example
> reg
===== ARM registers
(0) r0 (/32): 0x0000D3C2 (dirty)
(1) r1 (/32): 0xFD61F31C
(2) r2 (/32)
...
(164) ETM_contextid_comparator_mask (/32)
>
@end example
@end deffn

@deffn Command halt [ms]
@deffnx Command wait_halt [ms]
The @command{halt} command first sends a halt request to the target,
which @command{wait_halt} doesn't.
Otherwise these behave the same:  wait up to @var{ms} milliseconds,
or 5 seconds if there is no parameter, for the target to halt
(and enter debug mode).
Using 0 as the @var{ms} parameter prevents OpenOCD from waiting.

@quotation Warning
On ARM cores, software using the @emph{wait for interrupt} operation
often blocks the JTAG access needed by a @command{halt} command.
This is because that operation also puts the core into a low
power mode by gating the core clock;
but the core clock is needed to detect JTAG clock transitions.

One partial workaround uses adaptive clocking:  when the core is
interrupted the operation completes, then JTAG clocks are accepted
at least until the interrupt handler completes.
However, this workaround is often unusable since the processor, board,
and JTAG adapter must all support adaptive JTAG clocking.
Also, it can't work until an interrupt is issued.

A more complete workaround is to not use that operation while you
work with a JTAG debugger.
Tasking environments generaly have idle loops where the body is the
@emph{wait for interrupt} operation.
(On older cores, it is a coprocessor action;
newer cores have a @option{wfi} instruction.)
Such loops can just remove that operation, at the cost of higher
power consumption (because the CPU is needlessly clocked).
@end quotation

@end deffn

@deffn Command resume [address]
Resume the target at its current code position,
or the optional @var{address} if it is provided.
OpenOCD will wait 5 seconds for the target to resume.
@end deffn

@deffn Command step [address]
Single-step the target at its current code position,
or the optional @var{address} if it is provided.
@end deffn

@anchor{Reset Command}
@deffn Command reset
@deffnx Command {reset run}
@deffnx Command {reset halt}
@deffnx Command {reset init}
Perform as hard a reset as possible, using SRST if possible.
@emph{All defined targets will be reset, and target
events will fire during the reset sequence.}

The optional parameter specifies what should
happen after the reset.
If there is no parameter, a @command{reset run} is executed.
The other options will not work on all systems.
@xref{Reset Configuration}.

@itemize @minus
@item @b{run} Let the target run
@item @b{halt} Immediately halt the target
@item @b{init} Immediately halt the target, and execute the reset-init script
@end itemize
@end deffn

@deffn Command soft_reset_halt
Requesting target halt and executing a soft reset. This is often used
when a target cannot be reset and halted. The target, after reset is
released begins to execute code. OpenOCD attempts to stop the CPU and
then sets the program counter back to the reset vector. Unfortunately
the code that was executed may have left the hardware in an unknown
state.
@end deffn

@section I/O Utilities

These commands are available when
OpenOCD is built with @option{--enable-ioutil}.
They are mainly useful on embedded targets,
notably the ZY1000.
Hosts with operating systems have complementary tools.

@emph{Note:} there are several more such commands.

@deffn Command append_file filename [string]*
Appends the @var{string} parameters to
the text file @file{filename}.
Each string except the last one is followed by one space.
The last string is followed by a newline.
@end deffn

@deffn Command cat filename
Reads and displays the text file @file{filename}.
@end deffn

@deffn Command cp src_filename dest_filename
Copies contents from the file @file{src_filename}
into @file{dest_filename}.
@end deffn

@deffn Command ip
@emph{No description provided.}
@end deffn

@deffn Command ls
@emph{No description provided.}
@end deffn

@deffn Command mac
@emph{No description provided.}
@end deffn

@deffn Command  meminfo
Display available RAM memory on OpenOCD host.
Used in OpenOCD regression testing scripts.
@end deffn

@deffn Command peek
@emph{No description provided.}
@end deffn

@deffn Command poke
@emph{No description provided.}
@end deffn

@deffn Command rm filename
@c "rm" has both normal and Jim-level versions??
Unlinks the file @file{filename}.
@end deffn

@deffn Command trunc filename
Removes all data in the file @file{filename}.
@end deffn

@anchor{Memory access}
@section Memory access commands
@cindex memory access

These commands allow accesses of a specific size to the memory
system. Often these are used to configure the current target in some
special way. For example - one may need to write certain values to the
SDRAM controller to enable SDRAM.

@enumerate
@item Use the @command{targets} (plural) command
to change the current target.
@item In system level scripts these commands are deprecated.
Please use their TARGET object siblings to avoid making assumptions
about what TAP is the current target, or about MMU configuration.
@end enumerate

@deffn Command mdw [phys] addr [count]
@deffnx Command mdh [phys] addr [count]
@deffnx Command mdb [phys] addr [count]
Display contents of address @var{addr}, as
32-bit words (@command{mdw}), 16-bit halfwords (@command{mdh}),
or 8-bit bytes (@command{mdb}).
When the current target has an MMU which is present and active,
@var{addr} is interpreted as a virtual address.
Otherwise, or if the optional @var{phys} flag is specified,
@var{addr} is interpreted as a physical address.
If @var{count} is specified, displays that many units.
(If you want to manipulate the data instead of displaying it,
see the @code{mem2array} primitives.)
@end deffn

@deffn Command mww [phys] addr word
@deffnx Command mwh [phys] addr halfword
@deffnx Command mwb [phys] addr byte
Writes the specified @var{word} (32 bits),
@var{halfword} (16 bits), or @var{byte} (8-bit) value,
at the specified address @var{addr}.
When the current target has an MMU which is present and active,
@var{addr} is interpreted as a virtual address.
Otherwise, or if the optional @var{phys} flag is specified,
@var{addr} is interpreted as a physical address.
@end deffn


@anchor{Image access}
@section Image loading commands
@cindex image loading
@cindex image dumping

@anchor{dump_image}
@deffn Command {dump_image} filename address size
Dump @var{size} bytes of target memory starting at @var{address} to the
binary file named @var{filename}.
@end deffn

@deffn Command {fast_load}
Loads an image stored in memory by @command{fast_load_image} to the
current target. Must be preceeded by fast_load_image.
@end deffn

@deffn Command {fast_load_image} filename address [@option{bin}|@option{ihex}|@option{elf}]
Normally you should be using @command{load_image} or GDB load. However, for
testing purposes or when I/O overhead is significant(OpenOCD running on an embedded
host), storing the image in memory and uploading the image to the target
can be a way to upload e.g. multiple debug sessions when the binary does not change.
Arguments are the same as @command{load_image}, but the image is stored in OpenOCD host
memory, i.e. does not affect target.  This approach is also useful when profiling
target programming performance as I/O and target programming can easily be profiled
separately.
@end deffn

@anchor{load_image}
@deffn Command {load_image} filename address [@option{bin}|@option{ihex}|@option{elf}]
Load image from file @var{filename} to target memory at @var{address}.
The file format may optionally be specified
(@option{bin}, @option{ihex}, or @option{elf})
@end deffn

@deffn Command {test_image} filename [address [@option{bin}|@option{ihex}|@option{elf}]]
Displays image section sizes and addresses
as if @var{filename} were loaded into target memory
starting at @var{address} (defaults to zero).
The file format may optionally be specified
(@option{bin}, @option{ihex}, or @option{elf})
@end deffn

@deffn Command {verify_image} filename address [@option{bin}|@option{ihex}|@option{elf}]
Verify @var{filename} against target memory starting at @var{address}.
The file format may optionally be specified
(@option{bin}, @option{ihex}, or @option{elf})
This will first attempt a comparison using a CRC checksum, if this fails it will try a binary compare.
@end deffn


@section Breakpoint and Watchpoint commands
@cindex breakpoint
@cindex watchpoint

CPUs often make debug modules accessible through JTAG, with
hardware support for a handful of code breakpoints and data
watchpoints.
In addition, CPUs almost always support software breakpoints.

@deffn Command {bp} [address len [@option{hw}]]
With no parameters, lists all active breakpoints.
Else sets a breakpoint on code execution starting
at @var{address} for @var{length} bytes.
This is a software breakpoint, unless @option{hw} is specified
in which case it will be a hardware breakpoint.

(@xref{arm9 vector_catch}, or @pxref{xscale vector_catch},
for similar mechanisms that do not consume hardware breakpoints.)
@end deffn

@deffn Command {rbp} address
Remove the breakpoint at @var{address}.
@end deffn

@deffn Command {rwp} address
Remove data watchpoint on @var{address}
@end deffn

@deffn Command {wp} [address len [(@option{r}|@option{w}|@option{a}) [value [mask]]]]
With no parameters, lists all active watchpoints.
Else sets a data watchpoint on data from @var{address} for @var{length} bytes.
The watch point is an "access" watchpoint unless
the @option{r} or @option{w} parameter is provided,
defining it as respectively a read or write watchpoint.
If a @var{value} is provided, that value is used when determining if
the watchpoint should trigger.  The value may be first be masked
using @var{mask} to mark ``don't care'' fields.
@end deffn

@section Misc Commands

@cindex profiling
@deffn Command {profile} seconds filename
Profiling samples the CPU's program counter as quickly as possible,
which is useful for non-intrusive stochastic profiling.
Saves up to 10000 sampines in @file{filename} using ``gmon.out'' format.
@end deffn

@deffn Command {version}
Displays a string identifying the version of this OpenOCD server.
@end deffn

@deffn Command {virt2phys} virtual_address
Requests the current target to map the specified @var{virtual_address}
to its corresponding physical address, and displays the result.
@end deffn

@node Architecture and Core Commands
@chapter Architecture and Core Commands
@cindex Architecture Specific Commands
@cindex Core Specific Commands

Most CPUs have specialized JTAG operations to support debugging.
OpenOCD packages most such operations in its standard command framework.
Some of those operations don't fit well in that framework, so they are
exposed here as architecture or implementation (core) specific commands.

@anchor{ARM Hardware Tracing}
@section ARM Hardware Tracing
@cindex tracing
@cindex ETM
@cindex ETB

CPUs based on ARM cores may include standard tracing interfaces,
based on an ``Embedded Trace Module'' (ETM) which sends voluminous
address and data bus trace records to a ``Trace Port''.

@itemize
@item
Development-oriented boards will sometimes provide a high speed
trace connector for collecting that data, when the particular CPU
supports such an interface.
(The standard connector is a 38-pin Mictor, with both JTAG
and trace port support.)
Those trace connectors are supported by higher end JTAG adapters
and some logic analyzer modules; frequently those modules can
buffer several megabytes of trace data.
Configuring an ETM coupled to such an external trace port belongs
in the board-specific configuration file.
@item
If the CPU doesn't provide an external interface, it probably
has an ``Embedded Trace Buffer'' (ETB) on the chip, which is a
dedicated SRAM.  4KBytes is one common ETB size.
Configuring an ETM coupled only to an ETB belongs in the CPU-specific
(target) configuration file, since it works the same on all boards.
@end itemize

ETM support in OpenOCD doesn't seem to be widely used yet.

@quotation Issues
ETM support may be buggy, and at least some @command{etm config}
parameters should be detected by asking the ETM for them.

ETM trigger events could also implement a kind of complex
hardware breakpoint, much more powerful than the simple
watchpoint hardware exported by EmbeddedICE modules.
@emph{Such breakpoints can be triggered even when using the
dummy trace port driver}.

It seems like a GDB hookup should be possible,
as well as tracing only during specific states
(perhaps @emph{handling IRQ 23} or @emph{calls foo()}).

There should be GUI tools to manipulate saved trace data and help
analyse it in conjunction with the source code.
It's unclear how much of a common interface is shared
with the current XScale trace support, or should be
shared with eventual Nexus-style trace module support.

At this writing (November 2009) only ARM7, ARM9, and ARM11 support
for ETM modules is available.  The code should be able to
work with some newer cores; but not all of them support
this original style of JTAG access.
@end quotation

@subsection ETM Configuration
ETM setup is coupled with the trace port driver configuration.

@deffn {Config Command} {etm config} target width mode clocking driver
Declares the ETM associated with @var{target}, and associates it
with a given trace port @var{driver}.  @xref{Trace Port Drivers}.

Several of the parameters must reflect the trace port capabilities,
which are a function of silicon capabilties (exposed later
using @command{etm info}) and of what hardware is connected to
that port (such as an external pod, or ETB).
The @var{width} must be either 4, 8, or 16,
except with ETMv3.0 and newer modules which may also
support 1, 2, 24, 32, 48, and 64 bit widths.
(With those versions, @command{etm info} also shows whether
the selected port width and mode are supported.)

The @var{mode} must be @option{normal}, @option{multiplexed},
or @option{demultiplexed}.
The @var{clocking} must be @option{half} or @option{full}.

@quotation Warning
With ETMv3.0 and newer, the bits set with the @var{mode} and
@var{clocking} parameters both control the mode.
This modified mode does not map to the values supported by
previous ETM modules, so this syntax is subject to change.
@end quotation

@quotation Note
You can see the ETM registers using the @command{reg} command.
Not all possible registers are present in every ETM.
Most of the registers are write-only, and are used to configure
what CPU activities are traced.
@end quotation
@end deffn

@deffn Command {etm info}
Displays information about the current target's ETM.
This includes resource counts from the @code{ETM_CONFIG} register,
as well as silicon capabilities (except on rather old modules).
from the @code{ETM_SYS_CONFIG} register.
@end deffn

@deffn Command {etm status}
Displays status of the current target's ETM and trace port driver:
is the ETM idle, or is it collecting data?
Did trace data overflow?
Was it triggered?
@end deffn

@deffn Command {etm tracemode} [type context_id_bits cycle_accurate branch_output]
Displays what data that ETM will collect.
If arguments are provided, first configures that data.
When the configuration changes, tracing is stopped
and any buffered trace data is invalidated.

@itemize
@item @var{type} ... describing how data accesses are traced,
when they pass any ViewData filtering that that was set up.
The value is one of
@option{none} (save nothing),
@option{data} (save data),
@option{address} (save addresses),
@option{all} (save data and addresses)
@item @var{context_id_bits} ... 0, 8, 16, or 32
@item @var{cycle_accurate} ...  @option{enable} or @option{disable}
cycle-accurate instruction tracing.
Before ETMv3, enabling this causes much extra data to be recorded.
@item @var{branch_output} ...  @option{enable} or @option{disable}.
Disable this unless you need to try reconstructing the instruction
trace stream without an image of the code.
@end itemize
@end deffn

@deffn Command {etm trigger_debug} (@option{enable}|@option{disable})
Displays whether ETM triggering debug entry (like a breakpoint) is
enabled or disabled, after optionally modifying that configuration.
The default behaviour is @option{disable}.
Any change takes effect after the next @command{etm start}.

By using script commands to configure ETM registers, you can make the
processor enter debug state automatically when certain conditions,
more complex than supported by the breakpoint hardware, happen.
@end deffn

@subsection ETM Trace Operation

After setting up the ETM, you can use it to collect data.
That data can be exported to files for later analysis.
It can also be parsed with OpenOCD, for basic sanity checking.

To configure what is being traced, you will need to write
various trace registers using @command{reg ETM_*} commands.
For the definitions of these registers, read ARM publication
@emph{IHI 0014, ``Embedded Trace Macrocell, Architecture Specification''}.
Be aware that most of the relevant registers are write-only,
and that ETM resources are limited.  There are only a handful
of address comparators, data comparators, counters, and so on.

Examples of scenarios you might arrange to trace include:

@itemize
@item Code flow within a function, @emph{excluding} subroutines
it calls.  Use address range comparators to enable tracing
for instruction access within that function's body.
@item Code flow within a function, @emph{including} subroutines
it calls.  Use the sequencer and address comparators to activate
tracing on an ``entered function'' state, then deactivate it by
exiting that state when the function's exit code is invoked.
@item Code flow starting at the fifth invocation of a function,
combining one of the above models with a counter.
@item CPU data accesses to the registers for a particular device,
using address range comparators and the ViewData logic.
@item Such data accesses only during IRQ handling, combining the above
model with sequencer triggers which on entry and exit to the IRQ handler.
@item @emph{... more}
@end itemize

At this writing, September 2009, there are no Tcl utility
procedures to help set up any common tracing scenarios.

@deffn Command {etm analyze}
Reads trace data into memory, if it wasn't already present.
Decodes and prints the data that was collected.
@end deffn

@deffn Command {etm dump} filename
Stores the captured trace data in @file{filename}.
@end deffn

@deffn Command {etm image} filename [base_address] [type]
Opens an image file.
@end deffn

@deffn Command {etm load} filename
Loads captured trace data from @file{filename}.
@end deffn

@deffn Command {etm start}
Starts trace data collection.
@end deffn

@deffn Command {etm stop}
Stops trace data collection.
@end deffn

@anchor{Trace Port Drivers}
@subsection Trace Port Drivers

To use an ETM trace port it must be associated with a driver.

@deffn {Trace Port Driver} dummy
Use the @option{dummy} driver if you are configuring an ETM that's
not connected to anything (on-chip ETB or off-chip trace connector).
@emph{This driver lets OpenOCD talk to the ETM, but it does not expose
any trace data collection.}
@deffn {Config Command} {etm_dummy config} target
Associates the ETM for @var{target} with a dummy driver.
@end deffn
@end deffn

@deffn {Trace Port Driver} etb
Use the @option{etb} driver if you are configuring an ETM
to use on-chip ETB memory.
@deffn {Config Command} {etb config} target etb_tap
Associates the ETM for @var{target} with the ETB at @var{etb_tap}.
You can see the ETB registers using the @command{reg} command.
@end deffn
@deffn Command {etb trigger_percent} [percent]
This displays, or optionally changes, ETB behavior after the
ETM's configured @emph{trigger} event fires.
It controls how much more trace data is saved after the (single)
trace trigger becomes active.

@itemize
@item The default corresponds to @emph{trace around} usage,
recording 50 percent data before the event and the rest
afterwards.
@item The minimum value of @var{percent} is 2 percent,
recording almost exclusively data before the trigger.
Such extreme @emph{trace before} usage can help figure out
what caused that event to happen.
@item The maximum value of @var{percent} is 100 percent,
recording data almost exclusively after the event.
This extreme @emph{trace after} usage might help sort out
how the event caused trouble.
@end itemize
@c REVISIT allow "break" too -- enter debug mode.
@end deffn

@end deffn

@deffn {Trace Port Driver} oocd_trace
This driver isn't available unless OpenOCD was explicitly configured
with the @option{--enable-oocd_trace} option.  You probably don't want
to configure it unless you've built the appropriate prototype hardware;
it's @emph{proof-of-concept} software.

Use the @option{oocd_trace} driver if you are configuring an ETM that's
connected to an off-chip trace connector.

@deffn {Config Command} {oocd_trace config} target tty
Associates the ETM for @var{target} with a trace driver which
collects data through the serial port @var{tty}.
@end deffn

@deffn Command {oocd_trace resync}
Re-synchronizes with the capture clock.
@end deffn

@deffn Command {oocd_trace status}
Reports whether the capture clock is locked or not.
@end deffn
@end deffn


@section Generic ARM
@cindex ARM

These commands should be available on all ARM processors.
They are available in addition to other core-specific
commands that may be available.

@deffn Command {arm core_state} [@option{arm}|@option{thumb}]
Displays the core_state, optionally changing it to process
either @option{arm} or @option{thumb} instructions.
The target may later be resumed in the currently set core_state.
(Processors may also support the Jazelle state, but
that is not currently supported in OpenOCD.)
@end deffn

@deffn Command {arm disassemble} address [count [@option{thumb}]]
@cindex disassemble
Disassembles @var{count} instructions starting at @var{address}.
If @var{count} is not specified, a single instruction is disassembled.
If @option{thumb} is specified, or the low bit of the address is set,
Thumb2 (mixed 16/32-bit) instructions are used;
else ARM (32-bit) instructions are used.
(Processors may also support the Jazelle state, but
those instructions are not currently understood by OpenOCD.)

Note that all Thumb instructions are Thumb2 instructions,
so older processors (without Thumb2 support) will still
see correct disassembly of Thumb code.
Also, ThumbEE opcodes are the same as Thumb2,
with a handful of exceptions.
ThumbEE disassembly currently has no explicit support.
@end deffn

@deffn Command {arm mcr} pX op1 CRn CRm op2 value
Write @var{value} to a coprocessor @var{pX} register
passing parameters @var{CRn},
@var{CRm}, opcodes @var{opc1} and @var{opc2},
and using the MCR instruction.
(Parameter sequence matches the ARM instruction, but omits
an ARM register.)
@end deffn

@deffn Command {arm mrc} pX coproc op1 CRn CRm op2 
Read a coprocessor @var{pX} register passing parameters @var{CRn},
@var{CRm}, opcodes @var{opc1} and @var{opc2},
and the MRC instruction.
Returns the result so it can be manipulated by Jim scripts.
(Parameter sequence matches the ARM instruction, but omits
an ARM register.)
@end deffn

@deffn Command {arm reg}
Display a table of all banked core registers, fetching the current value from every
core mode if necessary.
@end deffn

@section ARMv4 and ARMv5 Architecture
@cindex ARMv4
@cindex ARMv5

The ARMv4 and ARMv5 architectures are widely used in embedded systems,
and introduced core parts of the instruction set in use today. 
That includes the Thumb instruction set, introduced in the ARMv4T
variant.

@subsection ARM7 and ARM9 specific commands
@cindex ARM7
@cindex ARM9

These commands are specific to ARM7 and ARM9 cores, like ARM7TDMI, ARM720T,
ARM9TDMI, ARM920T or ARM926EJ-S.
They are available in addition to the ARM commands,
and any other core-specific commands that may be available.

@deffn Command {arm7_9 dbgrq} (@option{enable}|@option{disable})
Control use of the EmbeddedIce DBGRQ signal to force entry into debug mode,
instead of breakpoints.  This should be
safe for all but ARM7TDMI--S cores (like Philips LPC).
This feature is enabled by default on most ARM9 cores,
including ARM9TDMI, ARM920T, and ARM926EJ-S.
@end deffn

@deffn Command {arm7_9 dcc_downloads} (@option{enable}|@option{disable})
@cindex DCC
Control the use of the debug communications channel (DCC) to write larger (>128 byte)
amounts of memory. DCC downloads offer a huge speed increase, but might be
unsafe, especially with targets running at very low speeds. This command was introduced
with OpenOCD rev. 60, and requires a few bytes of working area.
@end deffn

@anchor{arm7_9 fast_memory_access}
@deffn Command {arm7_9 fast_memory_access} (@option{enable}|@option{disable})
Enable or disable memory writes and reads that don't check completion of
the operation. This provides a huge speed increase, especially with USB JTAG
cables (FT2232), but might be unsafe if used with targets running at very low
speeds, like the 32kHz startup clock of an AT91RM9200.
@end deffn

@deffn Command {arm7_9 semihosting} [@option{enable}|@option{disable}]
@cindex ARM semihosting
Display status of semihosting, after optionally changing that status.

Semihosting allows for code executing on an ARM target to use the
I/O facilities on the host computer i.e. the system where OpenOCD
is running. The target application must be linked against a library
implementing the ARM semihosting convention that forwards operation
requests by using a special SVC instruction that is trapped at the
Supervisor Call vector by OpenOCD.
@end deffn

@subsection ARM720T specific commands
@cindex ARM720T

These commands are available to ARM720T based CPUs,
which are implementations of the ARMv4T architecture
based on the ARM7TDMI-S integer core.
They are available in addition to the ARM and ARM7/ARM9 commands.

@deffn Command {arm720t cp15} regnum [value]
Display cp15 register @var{regnum};
else if a @var{value} is provided, that value is written to that register.
@end deffn

@subsection ARM9 specific commands
@cindex ARM9

ARM9-family cores are built around ARM9TDMI or ARM9E (including ARM9EJS)
integer processors.
Such cores include the ARM920T, ARM926EJ-S, and ARM966.

@c 9-june-2009:  tried this on arm920t, it didn't work.
@c no-params always lists nothing caught, and that's how it acts.
@c 23-oct-2009:  doesn't work _consistently_ ... as if the ICE
@c versions have different rules about when they commit writes.

@anchor{arm9 vector_catch}
@deffn Command {arm9 vector_catch} [@option{all}|@option{none}|list]
@cindex vector_catch
Vector Catch hardware provides a sort of dedicated breakpoint
for hardware events such as reset, interrupt, and abort.
You can use this to conserve normal breakpoint resources,
so long as you're not concerned with code that branches directly
to those hardware vectors.

This always finishes by listing the current configuration.
If parameters are provided, it first reconfigures the
vector catch hardware to intercept
@option{all} of the hardware vectors,
@option{none} of them,
or a list with one or more of the following:
@option{reset} @option{undef} @option{swi} @option{pabt} @option{dabt}
@option{irq} @option{fiq}.
@end deffn

@subsection ARM920T specific commands
@cindex ARM920T

These commands are available to ARM920T based CPUs,
which are implementations of the ARMv4T architecture
built using the ARM9TDMI integer core.
They are available in addition to the ARM, ARM7/ARM9,
and ARM9 commands.

@deffn Command {arm920t cache_info}
Print information about the caches found. This allows to see whether your target
is an ARM920T (2x16kByte cache) or ARM922T (2x8kByte cache).
@end deffn

@deffn Command {arm920t cp15} regnum [value]
Display cp15 register @var{regnum};
else if a @var{value} is provided, that value is written to that register.
@end deffn

@deffn Command {arm920t cp15i} opcode [value [address]]
Interpreted access using cp15 @var{opcode}.
If no @var{value} is provided, the result is displayed.
Else if that value is written using the specified @var{address},
or using zero if no other address is not provided.
@end deffn

@deffn Command {arm920t read_cache} filename
Dump the content of ICache and DCache to a file named @file{filename}.
@end deffn

@deffn Command {arm920t read_mmu} filename
Dump the content of the ITLB and DTLB to a file named @file{filename}.
@end deffn

@subsection ARM926ej-s specific commands
@cindex ARM926ej-s

These commands are available to ARM926ej-s based CPUs,
which are implementations of the ARMv5TEJ architecture
based on the ARM9EJ-S integer core.
They are available in addition to the ARM, ARM7/ARM9,
and ARM9 commands.

The Feroceon cores also support these commands, although
they are not built from ARM926ej-s designs.

@deffn Command {arm926ejs cache_info}
Print information about the caches found.
@end deffn

@subsection ARM966E specific commands
@cindex ARM966E

These commands are available to ARM966 based CPUs,
which are implementations of the ARMv5TE architecture.
They are available in addition to the ARM, ARM7/ARM9,
and ARM9 commands.

@deffn Command {arm966e cp15} regnum [value]
Display cp15 register @var{regnum};
else if a @var{value} is provided, that value is written to that register.
@end deffn

@subsection XScale specific commands
@cindex XScale

Some notes about the debug implementation on the XScale CPUs:

The XScale CPU provides a special debug-only mini-instruction cache
(mini-IC) in which exception vectors and target-resident debug handler
code are placed by OpenOCD. In order to get access to the CPU, OpenOCD
must point vector 0 (the reset vector) to the entry of the debug
handler. However, this means that the complete first cacheline in the
mini-IC is marked valid, which makes the CPU fetch all exception
handlers from the mini-IC, ignoring the code in RAM.

OpenOCD currently does not sync the mini-IC entries with the RAM
contents (which would fail anyway while the target is running), so
the user must provide appropriate values using the @code{xscale
vector_table} command.

It is recommended to place a pc-relative indirect branch in the vector
table, and put the branch destination somewhere in memory. Doing so
makes sure the code in the vector table stays constant regardless of
code layout in memory:
@example
_vectors:
        ldr     pc,[pc,#0x100-8]
        ldr     pc,[pc,#0x100-8]
        ldr     pc,[pc,#0x100-8]
        ldr     pc,[pc,#0x100-8]
        ldr     pc,[pc,#0x100-8]
        ldr     pc,[pc,#0x100-8]
        ldr     pc,[pc,#0x100-8]
        ldr     pc,[pc,#0x100-8]
        .org 0x100
        .long real_reset_vector
        .long real_ui_handler
        .long real_swi_handler
        .long real_pf_abort
        .long real_data_abort
        .long 0 /* unused */
        .long real_irq_handler
        .long real_fiq_handler
@end example

The debug handler must be placed somewhere in the address space using
the @code{xscale debug_handler} command.  The allowed locations for the
debug handler are either (0x800 - 0x1fef800) or (0xfe000800 -
0xfffff800). The default value is 0xfe000800.


These commands are available to XScale based CPUs,
which are implementations of the ARMv5TE architecture.

@deffn Command {xscale analyze_trace}
Displays the contents of the trace buffer.
@end deffn

@deffn Command {xscale cache_clean_address} address
Changes the address used when cleaning the data cache.
@end deffn

@deffn Command {xscale cache_info}
Displays information about the CPU caches.
@end deffn

@deffn Command {xscale cp15} regnum [value]
Display cp15 register @var{regnum};
else if a @var{value} is provided, that value is written to that register.
@end deffn

@deffn Command {xscale debug_handler} target address
Changes the address used for the specified target's debug handler.
@end deffn

@deffn Command {xscale dcache} (@option{enable}|@option{disable})
Enables or disable the CPU's data cache.
@end deffn

@deffn Command {xscale dump_trace} filename
Dumps the raw contents of the trace buffer to @file{filename}.
@end deffn

@deffn Command {xscale icache} (@option{enable}|@option{disable})
Enables or disable the CPU's instruction cache.
@end deffn

@deffn Command {xscale mmu} (@option{enable}|@option{disable})
Enables or disable the CPU's memory management unit.
@end deffn

@deffn Command {xscale trace_buffer} (@option{enable}|@option{disable}) [@option{fill} [n] | @option{wrap}]
Enables or disables the trace buffer,
and controls how it is emptied.
@end deffn

@deffn Command {xscale trace_image} filename [offset [type]]
Opens a trace image from @file{filename}, optionally rebasing
its segment addresses by @var{offset}.
The image @var{type} may be one of
@option{bin} (binary), @option{ihex} (Intel hex),
@option{elf} (ELF file), @option{s19} (Motorola s19),
@option{mem}, or @option{builder}.
@end deffn

@anchor{xscale vector_catch}
@deffn Command {xscale vector_catch} [mask]
@cindex vector_catch
Display a bitmask showing the hardware vectors to catch.
If the optional parameter is provided, first set the bitmask to that value.

The mask bits correspond with bit 16..23 in the DCSR:
@example
0x01    Trap Reset
0x02    Trap Undefined Instructions
0x04    Trap Software Interrupt
0x08    Trap Prefetch Abort
0x10    Trap Data Abort
0x20    reserved
0x40    Trap IRQ
0x80    Trap FIQ
@end example
@end deffn

@anchor{xscale vector_table}
@deffn Command {xscale vector_table} [<low|high> <index> <value>]
@cindex vector_table

Set an entry in the mini-IC vector table. There are two tables: one for
low vectors (at 0x00000000), and one for high vectors (0xFFFF0000), each
holding the 8 exception vectors. @var{index} can be 1-7, because vector 0
points to the debug handler entry and can not be overwritten.
@var{value} holds the 32-bit opcode that is placed in the mini-IC.

Without arguments, the current settings are displayed.

@end deffn

@section ARMv6 Architecture
@cindex ARMv6

@subsection ARM11 specific commands
@cindex ARM11

@deffn Command {arm11 memwrite burst} [value]
Displays the value of the memwrite burst-enable flag,
which is enabled by default. Burst writes are only used
for memory writes larger than 1 word. Single word writes
are likely to be from reset init scripts and those writes
are often to non-memory locations which could easily have
many wait states, which could easily break burst writes.
If @var{value} is defined, first assigns that.
@end deffn

@deffn Command {arm11 memwrite error_fatal} [value]
Displays the value of the memwrite error_fatal flag,
which is enabled by default.
If @var{value} is defined, first assigns that.
@end deffn

@deffn Command {arm11 step_irq_enable}  [value]
Displays the value of the flag controlling whether
IRQs are enabled during single stepping;
they are disabled by default.
If @var{value} is defined, first assigns that.
@end deffn

@deffn Command {arm11 vcr} [value]
@cindex vector_catch
Displays the value of the @emph{Vector Catch Register (VCR)},
coprocessor 14 register 7.
If @var{value} is defined, first assigns that.

Vector Catch hardware provides dedicated breakpoints
for certain hardware events.
The specific bit values are core-specific (as in fact is using
coprocessor 14 register 7 itself) but all current ARM11
cores @emph{except the ARM1176} use the same six bits.
@end deffn

@section ARMv7 Architecture
@cindex ARMv7

@subsection ARMv7 Debug Access Port (DAP) specific commands
@cindex Debug Access Port
@cindex DAP
These commands are specific to ARM architecture v7 Debug Access Port (DAP),
included on Cortex-M3 and Cortex-A8 systems.
They are available in addition to other core-specific commands that may be available.

@deffn Command {dap info} [num]
Displays dap info for ap @var{num}, defaulting to the currently selected AP.
@end deffn

@deffn Command {dap apsel} [num]
Select AP @var{num}, defaulting to 0.
@end deffn

@deffn Command {dap apid} [num]
Displays id register from AP @var{num},
defaulting to the currently selected AP.
@end deffn

@deffn Command {dap baseaddr} [num]
Displays debug base address from AP @var{num},
defaulting to the currently selected AP.
@end deffn

@deffn Command {dap memaccess} [value]
Displays the number of extra tck for mem-ap memory bus access [0-255].
If @var{value} is defined, first assigns that.
@end deffn

@subsection Cortex-M3 specific commands
@cindex Cortex-M3

@deffn Command {cortex_m3 disassemble} address [count]
@cindex disassemble
Disassembles @var{count} Thumb2 instructions starting at @var{address}.
If @var{count} is not specified, a single instruction is disassembled.
@end deffn

@deffn Command {cortex_m3 maskisr} (@option{on}|@option{off})
Control masking (disabling) interrupts during target step/resume.
@end deffn

@deffn Command {cortex_m3 vector_catch} [@option{all}|@option{none}|list]
@cindex vector_catch
Vector Catch hardware provides dedicated breakpoints
for certain hardware events.

Parameters request interception of
@option{all} of these hardware event vectors,
@option{none} of them,
or one or more of the following:
@option{hard_err} for a HardFault exception;
@option{mm_err} for a MemManage exception;
@option{bus_err} for a BusFault exception;
@option{irq_err},
@option{state_err},
@option{chk_err}, or
@option{nocp_err} for various UsageFault exceptions; or
@option{reset}.
If NVIC setup code does not enable them,
MemManage, BusFault, and UsageFault exceptions
are mapped to HardFault.
UsageFault checks for
divide-by-zero and unaligned access
must also be explicitly enabled.

This finishes by listing the current vector catch configuration.
@end deffn

@anchor{Software Debug Messages and Tracing}
@section Software Debug Messages and Tracing
@cindex Linux-ARM DCC support
@cindex tracing
@cindex libdcc
@cindex DCC
OpenOCD can process certain requests from target software, when
the target uses appropriate libraries.
The most powerful mechanism is semihosting, but there is also
a lighter weight mechanism using only the DCC channel.

Currently @command{target_request debugmsgs}
is supported only for @option{arm7_9} and @option{cortex_m3} cores.
These messages are received as part of target polling, so
you need to have @command{poll on} active to receive them.
They are intrusive in that they will affect program execution
times.  If that is a problem, @pxref{ARM Hardware Tracing}.

See @file{libdcc} in the contrib dir for more details.
In addition to sending strings, characters, and
arrays of various size integers from the target,
@file{libdcc} also exports a software trace point mechanism.
The target being debugged may
issue trace messages which include a 24-bit @dfn{trace point} number.
Trace point support includes two distinct mechanisms,
each supported by a command:

@itemize
@item @emph{History} ... A circular buffer of trace points
can be set up, and then displayed at any time.
This tracks where code has been, which can be invaluable in
finding out how some fault was triggered.

The buffer may overflow, since it collects records continuously.
It may be useful to use some of the 24 bits to represent a
particular event, and other bits to hold data.

@item @emph{Counting} ... An array of counters can be set up,
and then displayed at any time.
This can help establish code coverage and identify hot spots.

The array of counters is directly indexed by the trace point
number, so trace points with higher numbers are not counted.
@end itemize

Linux-ARM kernels have a ``Kernel low-level debugging
via EmbeddedICE DCC channel'' option (CONFIG_DEBUG_ICEDCC,
depends on CONFIG_DEBUG_LL) which uses this mechanism to
deliver messages before a serial console can be activated.
This is not the same format used by @file{libdcc}.
Other software, such as the U-Boot boot loader, sometimes
does the same thing.

@deffn Command {target_request debugmsgs} [@option{enable}|@option{disable}|@option{charmsg}]
Displays current handling of target DCC message requests.
These messages may be sent to the debugger while the target is running.
The optional @option{enable} and @option{charmsg} parameters
both enable the messages, while @option{disable} disables them.

With @option{charmsg} the DCC words each contain one character,
as used by Linux with CONFIG_DEBUG_ICEDCC;
otherwise the libdcc format is used.
@end deffn

@deffn Command {trace history} [@option{clear}|count]
With no parameter, displays all the trace points that have triggered
in the order they triggered.
With the parameter @option{clear}, erases all current trace history records.
With a @var{count} parameter, allocates space for that many
history records.
@end deffn

@deffn Command {trace point} [@option{clear}|identifier]
With no parameter, displays all trace point identifiers and how many times
they have been triggered.
With the parameter @option{clear}, erases all current trace point counters.
With a numeric @var{identifier} parameter, creates a new a trace point counter
and associates it with that identifier.

@emph{Important:}  The identifier and the trace point number
are not related except by this command.
These trace point numbers always start at zero (from server startup,
or after @command{trace point clear}) and count up from there.
@end deffn


@node JTAG Commands
@chapter JTAG Commands
@cindex JTAG Commands
Most general purpose JTAG commands have been presented earlier.
(@xref{JTAG Speed}, @ref{Reset Configuration}, and @ref{TAP Declaration}.)
Lower level JTAG commands, as presented here,
may be needed to work with targets which require special
attention during operations such as reset or initialization.

To use these commands you will need to understand some
of the basics of JTAG, including:

@itemize @bullet
@item A JTAG scan chain consists of a sequence of individual TAP
devices such as a CPUs.
@item Control operations involve moving each TAP through the same
standard state machine (in parallel)
using their shared TMS and clock signals.
@item Data transfer involves shifting data through the chain of
instruction or data registers of each TAP, writing new register values
while the reading previous ones.
@item Data register sizes are a function of the instruction active in
a given TAP, while instruction register sizes are fixed for each TAP.
All TAPs support a BYPASS instruction with a single bit data register.
@item The way OpenOCD differentiates between TAP devices is by
shifting different instructions into (and out of) their instruction
registers.
@end itemize

@section Low Level JTAG Commands

These commands are used by developers who need to access
JTAG instruction or data registers, possibly controlling
the order of TAP state transitions.
If you're not debugging OpenOCD internals, or bringing up a
new JTAG adapter or a new type of TAP device (like a CPU or
JTAG router), you probably won't need to use these commands.

@deffn Command {drscan} tap [numbits value]+ [@option{-endstate} tap_state]
Loads the data register of @var{tap} with a series of bit fields
that specify the entire register.
Each field is @var{numbits} bits long with
a numeric @var{value} (hexadecimal encouraged).
The return value holds the original value of each
of those fields.

For example, a 38 bit number might be specified as one
field of 32 bits then one of 6 bits.
@emph{For portability, never pass fields which are more
than 32 bits long.  Many OpenOCD implementations do not
support 64-bit (or larger) integer values.}

All TAPs other than @var{tap} must be in BYPASS mode.
The single bit in their data registers does not matter.

When @var{tap_state} is specified, the JTAG state machine is left
in that state.
For example @sc{drpause} might be specified, so that more
instructions can be issued before re-entering the @sc{run/idle} state.
If the end state is not specified, the @sc{run/idle} state is entered.

@quotation Warning
OpenOCD does not record information about data register lengths,
so @emph{it is important that you get the bit field lengths right}.
Remember that different JTAG instructions refer to different
data registers, which may have different lengths.
Moreover, those lengths may not be fixed;
the SCAN_N instruction can change the length of
the register accessed by the INTEST instruction
(by connecting a different scan chain).
@end quotation
@end deffn

@deffn Command {flush_count}
Returns the number of times the JTAG queue has been flushed.
This may be used for performance tuning.

For example, flushing a queue over USB involves a
minimum latency, often several milliseconds, which does
not change with the amount of data which is written.
You may be able to identify performance problems by finding
tasks which waste bandwidth by flushing small transfers too often,
instead of batching them into larger operations.
@end deffn

@deffn Command {irscan} [tap instruction]+ [@option{-endstate} tap_state]
For each @var{tap} listed, loads the instruction register
with its associated numeric @var{instruction}.
(The number of bits in that instruction may be displayed
using the @command{scan_chain} command.)
For other TAPs, a BYPASS instruction is loaded.

When @var{tap_state} is specified, the JTAG state machine is left
in that state.
For example @sc{irpause} might be specified, so the data register
can be loaded before re-entering the @sc{run/idle} state.
If the end state is not specified, the @sc{run/idle} state is entered.

@quotation Note
OpenOCD currently supports only a single field for instruction
register values, unlike data register values.
For TAPs where the instruction register length is more than 32 bits,
portable scripts currently must issue only BYPASS instructions.
@end quotation
@end deffn

@deffn Command {jtag_reset} trst srst
Set values of reset signals.
The @var{trst} and @var{srst} parameter values may be
@option{0}, indicating that reset is inactive (pulled or driven high),
or @option{1}, indicating it is active (pulled or driven low).
The @command{reset_config} command should already have been used
to configure how the board and JTAG adapter treat these two
signals, and to say if either signal is even present.
@xref{Reset Configuration}.

Note that TRST is specially handled.
It actually signifies JTAG's @sc{reset} state.
So if the board doesn't support the optional TRST signal,
or it doesn't support it along with the specified SRST value,
JTAG reset is triggered with TMS and TCK signals
instead of the TRST signal.
And no matter how that JTAG reset is triggered, once
the scan chain enters @sc{reset} with TRST inactive,
TAP @code{post-reset} events are delivered to all TAPs
with handlers for that event.
@end deffn

@deffn Command {pathmove} start_state [next_state ...]
Start by moving to @var{start_state}, which
must be one of the @emph{stable} states.
Unless it is the only state given, this will often be the
current state, so that no TCK transitions are needed.
Then, in a series of single state transitions
(conforming to the JTAG state machine) shift to
each @var{next_state} in sequence, one per TCK cycle.
The final state must also be stable.
@end deffn

@deffn Command {runtest} @var{num_cycles}
Move to the @sc{run/idle} state, and execute at least
@var{num_cycles} of the JTAG clock (TCK).
Instructions often need some time
to execute before they take effect.
@end deffn

@c tms_sequence (short|long)
@c ... temporary, debug-only, other than USBprog bug workaround...

@deffn Command {verify_ircapture} (@option{enable}|@option{disable})
Verify values captured during @sc{ircapture} and returned
during IR scans.  Default is enabled, but this can be
overridden by @command{verify_jtag}.
This flag is ignored when validating JTAG chain configuration.
@end deffn

@deffn Command {verify_jtag} (@option{enable}|@option{disable})
Enables verification of DR and IR scans, to help detect
programming errors.  For IR scans, @command{verify_ircapture}
must also be enabled.
Default is enabled.
@end deffn

@section TAP state names
@cindex TAP state names

The @var{tap_state} names used by OpenOCD in the @command{drscan},
@command{irscan}, and @command{pathmove} commands are the same
as those used in SVF boundary scan documents, except that
SVF uses @sc{idle} instead of @sc{run/idle}.

@itemize @bullet
@item @b{RESET} ... @emph{stable} (with TMS high);
acts as if TRST were pulsed
@item @b{RUN/IDLE} ... @emph{stable}; don't assume this always means IDLE
@item @b{DRSELECT}
@item @b{DRCAPTURE}
@item @b{DRSHIFT} ... @emph{stable}; TDI/TDO shifting
through the data register
@item @b{DREXIT1}
@item @b{DRPAUSE} ... @emph{stable}; data register ready
for update or more shifting
@item @b{DREXIT2}
@item @b{DRUPDATE}
@item @b{IRSELECT}
@item @b{IRCAPTURE}
@item @b{IRSHIFT} ... @emph{stable}; TDI/TDO shifting
through the instruction register
@item @b{IREXIT1}
@item @b{IRPAUSE} ... @emph{stable}; instruction register ready
for update or more shifting
@item @b{IREXIT2}
@item @b{IRUPDATE}
@end itemize

Note that only six of those states are fully ``stable'' in the
face of TMS fixed (low except for @sc{reset})
and a free-running JTAG clock.  For all the
others, the next TCK transition changes to a new state.

@itemize @bullet
@item From @sc{drshift} and @sc{irshift}, clock transitions will
produce side effects by changing register contents.  The values
to be latched in upcoming @sc{drupdate} or @sc{irupdate} states
may not be as expected.
@item @sc{run/idle}, @sc{drpause}, and @sc{irpause} are reasonable
choices after @command{drscan} or @command{irscan} commands,
since they are free of JTAG side effects.
@item @sc{run/idle} may have side effects that appear at non-JTAG
levels, such as advancing the ARM9E-S instruction pipeline.
Consult the documentation for the TAP(s) you are working with.
@end itemize

@node Boundary Scan Commands
@chapter Boundary Scan Commands

One of the original purposes of JTAG was to support
boundary scan based hardware testing.
Although its primary focus is to support On-Chip Debugging,
OpenOCD also includes some boundary scan commands.

@section SVF: Serial Vector Format
@cindex Serial Vector Format
@cindex SVF

The Serial Vector Format, better known as @dfn{SVF}, is a
way to represent JTAG test patterns in text files.
OpenOCD supports running such test files.

@deffn Command {svf} filename [@option{quiet}]
This issues a JTAG reset (Test-Logic-Reset) and then
runs the SVF script from @file{filename}.
Unless the @option{quiet} option is specified,
each command is logged before it is executed.
@end deffn

@section XSVF: Xilinx Serial Vector Format
@cindex Xilinx Serial Vector Format
@cindex XSVF

The Xilinx Serial Vector Format, better known as @dfn{XSVF}, is a
binary representation of SVF which is optimized for use with
Xilinx devices.
OpenOCD supports running such test files.

@quotation Important
Not all XSVF commands are supported.
@end quotation

@deffn Command {xsvf} (tapname|@option{plain}) filename [@option{virt2}] [@option{quiet}]
This issues a JTAG reset (Test-Logic-Reset) and then
runs the XSVF script from @file{filename}.
When a @var{tapname} is specified, the commands are directed at
that TAP.
When @option{virt2} is specified, the @sc{xruntest} command counts
are interpreted as TCK cycles instead of microseconds.
Unless the @option{quiet} option is specified,
messages are logged for comments and some retries.
@end deffn

The OpenOCD sources also include two utility scripts
for working with XSVF; they are not currently installed
after building the software.
You may find them useful:

@itemize
@item @emph{svf2xsvf} ... converts SVF files into the extended XSVF
syntax understood by the @command{xsvf} command; see notes below.
@item @emph{xsvfdump} ... converts XSVF files into a text output format;
understands the OpenOCD extensions.
@end itemize

The input format accepts a handful of non-standard extensions.
These include three opcodes corresponding to SVF extensions
from Lattice Semiconductor (LCOUNT, LDELAY, LDSR), and
two opcodes supporting a more accurate translation of SVF
(XTRST, XWAITSTATE).
If @emph{xsvfdump} shows a file is using those opcodes, it
probably will not be usable with other XSVF tools.


@node TFTP
@chapter TFTP
@cindex TFTP
If OpenOCD runs on an embedded host(as ZY1000 does), then TFTP can
be used to access files on PCs (either the developer's PC or some other PC).

The way this works on the ZY1000 is to prefix a filename by
"/tftp/ip/" and append the TFTP path on the TFTP
server (tftpd). For example,

@example
load_image /tftp/10.0.0.96/c:\temp\abc.elf
@end example

will load c:\temp\abc.elf from the developer pc (10.0.0.96) into memory as
if the file was hosted on the embedded host.

In order to achieve decent performance, you must choose a TFTP server
that supports a packet size bigger than the default packet size (512 bytes). There
are numerous TFTP servers out there (free and commercial) and you will have to do
a bit of googling to find something that fits your requirements.

@node GDB and OpenOCD
@chapter GDB and OpenOCD
@cindex GDB
OpenOCD complies with the remote gdbserver protocol, and as such can be used
to debug remote targets.
Setting up GDB to work with OpenOCD can involve several components:

@itemize
@item OpenOCD itself may need to be configured.  @xref{GDB Configuration}.
@item GDB itself may need configuration, as shown in this chapter.
@item If you have a GUI environment like Eclipse,
that also will probably need to be configured.
@end itemize

Of course, the version of GDB you use will need to be one which has
been built to know about the target CPU you're using.  It's probably
part of the tool chain you're using.  For example, if you are doing
cross-development for ARM on an x86 PC, instead of using the native
x86 @command{gdb} command you might use @command{arm-none-eabi-gdb}
if that's the tool chain used to compile your code.

@anchor{Connecting to GDB}
@section Connecting to GDB
@cindex Connecting to GDB
Use GDB 6.7 or newer with OpenOCD if you run into trouble. For
instance GDB 6.3 has a known bug that produces bogus memory access
errors, which has since been fixed: look up 1836 in
@url{http://sourceware.org/cgi-bin/gnatsweb.pl?database=gdb}

OpenOCD can communicate with GDB in two ways:

@enumerate
@item
A socket (TCP/IP) connection is typically started as follows:
@example
target remote localhost:3333
@end example
This would cause GDB to connect to the gdbserver on the local pc using port 3333.
@item
A pipe connection is typically started as follows:
@example
target remote | openocd --pipe
@end example
This would cause GDB to run OpenOCD and communicate using pipes (stdin/stdout).
Using this method has the advantage of GDB starting/stopping OpenOCD for the debug
session.
@end enumerate

To list the available OpenOCD commands type @command{monitor help} on the
GDB command line.

@section Configuring GDB for OpenOCD

OpenOCD supports the gdb @option{qSupported} packet, this enables information
to be sent by the GDB remote server (i.e. OpenOCD) to GDB. Typical information includes
packet size and the device's memory map.
You do not need to configure the packet size by hand,
and the relevant parts of the memory map should be automatically
set up when you declare (NOR) flash banks.

However, there are other things which GDB can't currently query.
You may need to set those up by hand.
As OpenOCD starts up, you will often see a line reporting
something like:

@example
Info : lm3s.cpu: hardware has 6 breakpoints, 4 watchpoints
@end example

You can pass that information to GDB with these commands:

@example
set remote hardware-breakpoint-limit 6
set remote hardware-watchpoint-limit 4
@end example

With that particular hardware (Cortex-M3) the hardware breakpoints
only work for code running from flash memory.  Most other ARM systems
do not have such restrictions.

@section Programming using GDB
@cindex Programming using GDB

By default the target memory map is sent to GDB. This can be disabled by
the following OpenOCD configuration option:
@example
gdb_memory_map disable
@end example
For this to function correctly a valid flash configuration must also be set
in OpenOCD. For faster performance you should also configure a valid
working area.

Informing GDB of the memory map of the target will enable GDB to protect any
flash areas of the target and use hardware breakpoints by default. This means
that the OpenOCD option @command{gdb_breakpoint_override} is not required when
using a memory map. @xref{gdb_breakpoint_override}.

To view the configured memory map in GDB, use the GDB command @option{info mem}
All other unassigned addresses within GDB are treated as RAM.

GDB 6.8 and higher set any memory area not in the memory map as inaccessible.
This can be changed to the old behaviour by using the following GDB command
@example
set mem inaccessible-by-default off
@end example

If @command{gdb_flash_program enable} is also used, GDB will be able to
program any flash memory using the vFlash interface.

GDB will look at the target memory map when a load command is given, if any
areas to be programmed lie within the target flash area the vFlash packets
will be used.

If the target needs configuring before GDB programming, an event
script can be executed:
@example
$_TARGETNAME configure -event EVENTNAME BODY
@end example

To verify any flash programming the GDB command @option{compare-sections}
can be used.

@node Tcl Scripting API
@chapter Tcl Scripting API
@cindex Tcl Scripting API
@cindex Tcl scripts
@section API rules

The commands are stateless. E.g. the telnet command line has a concept
of currently active target, the Tcl API proc's take this sort of state
information as an argument to each proc.

There are three main types of return values: single value, name value
pair list and lists.

Name value pair. The proc 'foo' below returns a name/value pair
list.

@verbatim

 >  set foo(me)  Duane
 >  set foo(you) Oyvind
 >  set foo(mouse) Micky
 >  set foo(duck) Donald

If one does this:

 >  set foo

The result is:

    me Duane you Oyvind mouse Micky duck Donald

Thus, to get the names of the associative array is easy:

     foreach  { name value }   [set foo]   {
                puts "Name: $name, Value: $value"
     }
@end verbatim

Lists returned must be relatively small. Otherwise a range
should be passed in to the proc in question.

@section Internal low-level Commands

By low-level, the intent is a human would not directly use these commands.

Low-level commands are (should be) prefixed with "ocd_", e.g.
@command{ocd_flash_banks}
is the low level API upon which @command{flash banks} is implemented.

@itemize @bullet
@item @b{ocd_mem2array} <@var{varname}> <@var{width}> <@var{addr}> <@var{nelems}>

Read memory and return as a Tcl array for script processing
@item @b{ocd_array2mem} <@var{varname}> <@var{width}> <@var{addr}> <@var{nelems}>

Convert a Tcl array to memory locations and write the values
@item @b{ocd_flash_banks} <@var{driver}> <@var{base}> <@var{size}> <@var{chip_width}> <@var{bus_width}> <@var{target}> [@option{driver options} ...]

Return information about the flash banks
@end itemize

OpenOCD commands can consist of two words, e.g. "flash banks". The
@file{startup.tcl} "unknown" proc will translate this into a Tcl proc
called "flash_banks".

@section OpenOCD specific Global Variables

Real Tcl has ::tcl_platform(), and platform::identify, and many other
variables. JimTCL, as implemented in OpenOCD creates $ocd_HOSTOS which
holds one of the following values:

@itemize @bullet
@item @b{winxx}    Built using Microsoft Visual Studio
@item @b{linux}    Linux is the underlying operating sytem
@item @b{darwin}   Darwin (mac-os) is the underlying operating sytem.
@item @b{cygwin}   Running under Cygwin
@item @b{mingw32}  Running under MingW32
@item @b{other}    Unknown, none of the above.
@end itemize

Note: 'winxx' was choosen because today (March-2009) no distinction is made between Win32 and Win64.

@quotation Note
We should add support for a variable like Tcl variable
@code{tcl_platform(platform)}, it should be called
@code{jim_platform} (because it
is jim, not real tcl).
@end quotation

@node FAQ
@chapter FAQ
@cindex faq
@enumerate
@anchor{FAQ RTCK}
@item @b{RTCK, also known as: Adaptive Clocking - What is it?}
@cindex RTCK
@cindex adaptive clocking
@*

In digital circuit design it is often refered to as ``clock
synchronisation'' the JTAG interface uses one clock (TCK or TCLK)
operating at some speed, your target is operating at another.  The two
clocks are not synchronised, they are ``asynchronous''

In order for the two to work together they must be synchronised. Otherwise
the two systems will get out of sync with each other and nothing will
work. There are 2 basic options:
@enumerate
@item
Use a special circuit.
@item
One clock must be some multiple slower than the other.
@end enumerate

@b{Does this really matter?} For some chips and some situations, this
is a non-issue (i.e.: A 500MHz ARM926) but for others - for example some
Atmel SAM7 and SAM9 chips start operation from reset at 32kHz -
program/enable the oscillators and eventually the main clock. It is in
those critical times you must slow the JTAG clock to sometimes 1 to
4kHz.

Imagine debugging a 500MHz ARM926 hand held battery powered device
that ``deep sleeps'' at 32kHz between every keystroke. It can be
painful.

@b{Solution #1 - A special circuit}

In order to make use of this, your JTAG dongle must support the RTCK
feature. Not all dongles support this - keep reading!

The RTCK signal often found in some ARM chips is used to help with
this problem. ARM has a good description of the problem described at
this link: @url{http://www.arm.com/support/faqdev/4170.html} [checked
28/nov/2008]. Link title: ``How does the JTAG synchronisation logic
work? / how does adaptive clocking work?''.

The nice thing about adaptive clocking is that ``battery powered hand
held device example'' - the adaptiveness works perfectly all the
time. One can set a break point or halt the system in the deep power
down code, slow step out until the system speeds up.

Note that adaptive clocking may also need to work at the board level,
when a board-level scan chain has multiple chips.
Parallel clock voting schemes are good way to implement this,
both within and between chips, and can easily be implemented
with a CPLD.
It's not difficult to have logic fan a module's input TCK signal out
to each TAP in the scan chain, and then wait until each TAP's RTCK comes
back with the right polarity before changing the output RTCK signal.
Texas Instruments makes some clock voting logic available
for free (with no support) in VHDL form; see
@url{http://tiexpressdsp.com/index.php/Adaptive_Clocking}

@b{Solution #2 - Always works - but may be slower}

Often this is a perfectly acceptable solution.

In most simple terms: Often the JTAG clock must be 1/10 to 1/12 of
the target clock speed. But what that ``magic division'' is varies
depending on the chips on your board.
@b{ARM rule of thumb} Most ARM based systems require an 6:1 division;
ARM11 cores use an 8:1 division.
@b{Xilinx rule of thumb} is 1/12 the clock speed.

Note: Many FTDI2232C based JTAG dongles are limited to 6MHz.

You can still debug the 'low power' situations - you just need to
manually adjust the clock speed at every step. While painful and
tedious, it is not always practical.

It is however easy to ``code your way around it'' - i.e.: Cheat a little,
have a special debug mode in your application that does a ``high power
sleep''. If you are careful - 98% of your problems can be debugged
this way.

Note that on ARM you may need to avoid using the @emph{wait for interrupt}
operation in your idle loops even if you don't otherwise change the CPU
clock rate.
That operation gates the CPU clock, and thus the JTAG clock; which
prevents JTAG access.  One consequence is not being able to @command{halt}
cores which are executing that @emph{wait for interrupt} operation.

To set the JTAG frequency use the command:

@example
# Example: 1.234MHz
jtag_khz 1234
@end example


@item @b{Win32 Pathnames} Why don't backslashes work in Windows paths?

OpenOCD uses Tcl and a backslash is an escape char. Use @{ and @}
around Windows filenames.

@example
> echo \a

> echo @{\a@}
\a
> echo "\a"

>
@end example


@item @b{Missing: cygwin1.dll} OpenOCD complains about a missing cygwin1.dll.

Make sure you have Cygwin installed, or at least a version of OpenOCD that
claims to come with all the necessary DLLs. When using Cygwin, try launching
OpenOCD from the Cygwin shell.

@item @b{Breakpoint Issue} I'm trying to set a breakpoint using GDB (or a frontend like Insight or
Eclipse), but OpenOCD complains that "Info: arm7_9_common.c:213
arm7_9_add_breakpoint(): sw breakpoint requested, but software breakpoints not enabled".

GDB issues software breakpoints when a normal breakpoint is requested, or to implement
source-line single-stepping. On ARMv4T systems, like ARM7TDMI, ARM720T or ARM920T,
software breakpoints consume one of the two available hardware breakpoints.

@item @b{LPC2000 Flash} When erasing or writing LPC2000 on-chip flash, the operation fails at random.

Make sure the core frequency specified in the @option{flash lpc2000} line matches the
clock at the time you're programming the flash. If you've specified the crystal's
frequency, make sure the PLL is disabled. If you've specified the full core speed
(e.g. 60MHz), make sure the PLL is enabled.

@item @b{Amontec Chameleon} When debugging using an Amontec Chameleon in its JTAG Accelerator configuration,
I keep getting "Error: amt_jtagaccel.c:184 amt_wait_scan_busy(): amt_jtagaccel timed
out while waiting for end of scan, rtck was disabled".

Make sure your PC's parallel port operates in EPP mode. You might have to try several
settings in your PC BIOS (ECP, EPP, and different versions of those).

@item @b{Data Aborts} When debugging with OpenOCD and GDB (plain GDB, Insight, or Eclipse),
I get lots of "Error: arm7_9_common.c:1771 arm7_9_read_memory():
memory read caused data abort".

The errors are non-fatal, and are the result of GDB trying to trace stack frames
beyond the last valid frame. It might be possible to prevent this by setting up
a proper "initial" stack frame, if you happen to know what exactly has to
be done, feel free to add this here.

@b{Simple:} In your startup code - push 8 registers of zeros onto the
stack before calling main(). What GDB is doing is ``climbing'' the run
time stack by reading various values on the stack using the standard
call frame for the target. GDB keeps going - until one of 2 things
happen @b{#1} an invalid frame is found, or @b{#2} some huge number of
stackframes have been processed. By pushing zeros on the stack, GDB
gracefully stops.

@b{Debugging Interrupt Service Routines} - In your ISR before you call
your C code, do the same - artifically push some zeros onto the stack,
remember to pop them off when the ISR is done.

@b{Also note:} If you have a multi-threaded operating system, they
often do not @b{in the intrest of saving memory} waste these few
bytes. Painful...


@item @b{JTAG Reset Config} I get the following message in the OpenOCD console (or log file):
"Warning: arm7_9_common.c:679 arm7_9_assert_reset(): srst resets test logic, too".

This warning doesn't indicate any serious problem, as long as you don't want to
debug your core right out of reset. Your .cfg file specified @option{jtag_reset
trst_and_srst srst_pulls_trst} to tell OpenOCD that either your board,
your debugger or your target uC (e.g. LPC2000) can't assert the two reset signals
independently. With this setup, it's not possible to halt the core right out of
reset, everything else should work fine.

@item @b{USB Power} When using OpenOCD in conjunction with Amontec JTAGkey and the Yagarto
toolchain (Eclipse, arm-elf-gcc, arm-elf-gdb), the debugging seems to be
unstable. When single-stepping over large blocks of code, GDB and OpenOCD
quit with an error message. Is there a stability issue with OpenOCD?

No, this is not a stability issue concerning OpenOCD. Most users have solved
this issue by simply using a self-powered USB hub, which they connect their
Amontec JTAGkey to. Apparently, some computers do not provide a USB power
supply stable enough for the Amontec JTAGkey to be operated.

@b{Laptops running on battery have this problem too...}

@item @b{USB Power} When using the Amontec JTAGkey, sometimes OpenOCD crashes with the
following error messages: "Error: ft2232.c:201 ft2232_read(): FT_Read returned:
4" and "Error: ft2232.c:365 ft2232_send_and_recv(): couldn't read from FT2232".
What does that mean and what might be the reason for this?

First of all, the reason might be the USB power supply. Try using a self-powered
hub instead of a direct connection to your computer. Secondly, the error code 4
corresponds to an FT_IO_ERROR, which means that the driver for the FTDI USB
chip ran into some sort of error - this points us to a USB problem.

@item @b{GDB Disconnects} When using the Amontec JTAGkey, sometimes OpenOCD crashes with the following
error message: "Error: gdb_server.c:101 gdb_get_char(): read: 10054".
What does that mean and what might be the reason for this?

Error code 10054 corresponds to WSAECONNRESET, which means that the debugger (GDB)
has closed the connection to OpenOCD. This might be a GDB issue.

@item @b{LPC2000 Flash} In the configuration file in the section where flash device configurations
are described, there is a parameter for specifying the clock frequency
for LPC2000 internal flash devices (e.g.  @option{flash bank lpc2000
0x0 0x40000 0 0 0 lpc2000_v1 14746 calc_checksum}), which must be
specified in kilohertz. However, I do have a quartz crystal of a
frequency that contains fractions of kilohertz (e.g. 14,745,600 Hz,
i.e. 14,745.600 kHz).  Is it possible to specify real numbers for the
clock frequency?

No. The clock frequency specified here must be given as an integral number.
However, this clock frequency is used by the In-Application-Programming (IAP)
routines of the LPC2000 family only, which seems to be very tolerant concerning
the given clock frequency, so a slight difference between the specified clock
frequency and the actual clock frequency will not cause any trouble.

@item @b{Command Order} Do I have to keep a specific order for the commands in the configuration file?

Well, yes and no. Commands can be given in arbitrary order, yet the
devices listed for the JTAG scan chain must be given in the right
order (jtag newdevice), with the device closest to the TDO-Pin being
listed first. In general, whenever objects of the same type exist
which require an index number, then these objects must be given in the
right order (jtag newtap, targets and flash banks - a target
references a jtag newtap and a flash bank references a target).

You can use the ``scan_chain'' command to verify and display the tap order.

Also, some commands can't execute until after @command{init} has been
processed.  Such commands include @command{nand probe} and everything
else that needs to write to controller registers, perhaps for setting
up DRAM and loading it with code.

@anchor{FAQ TAP Order}
@item @b{JTAG TAP Order} Do I have to declare the TAPS in some
particular order?

Yes; whenever you have more than one, you must declare them in
the same order used by the hardware.

Many newer devices have multiple JTAG TAPs. For example: ST
Microsystems STM32 chips have two TAPs, a ``boundary scan TAP'' and
``Cortex-M3'' TAP.  Example: The STM32 reference manual, Document ID:
RM0008, Section 26.5, Figure 259, page 651/681, the ``TDI'' pin is
connected to the boundary scan TAP, which then connects to the
Cortex-M3 TAP, which then connects to the TDO pin.

Thus, the proper order for the STM32 chip is: (1) The Cortex-M3, then
(2) The boundary scan TAP. If your board includes an additional JTAG
chip in the scan chain (for example a Xilinx CPLD or FPGA) you could
place it before or after the STM32 chip in the chain. For example:

@itemize @bullet
@item OpenOCD_TDI(output) -> STM32 TDI Pin (BS Input)
@item STM32 BS TDO (output) -> STM32 Cortex-M3 TDI (input)
@item STM32 Cortex-M3 TDO (output) -> SM32 TDO Pin
@item STM32 TDO Pin (output) -> Xilinx TDI Pin (input)
@item Xilinx TDO Pin -> OpenOCD TDO (input)
@end itemize

The ``jtag device'' commands would thus be in the order shown below. Note:

@itemize @bullet
@item jtag newtap Xilinx tap -irlen ...
@item jtag newtap stm32  cpu -irlen ...
@item jtag newtap stm32  bs  -irlen ...
@item # Create the debug target and say where it is
@item target create stm32.cpu -chain-position stm32.cpu ...
@end itemize


@item @b{SYSCOMP} Sometimes my debugging session terminates with an error. When I look into the
log file, I can see these error messages: Error: arm7_9_common.c:561
arm7_9_execute_sys_speed(): timeout waiting for SYSCOMP

TODO.

@end enumerate

@node Tcl Crash Course
@chapter Tcl Crash Course
@cindex Tcl

Not everyone knows Tcl - this is not intended to be a replacement for
learning Tcl, the intent of this chapter is to give you some idea of
how the Tcl scripts work.

This chapter is written with two audiences in mind. (1) OpenOCD users
who need to understand a bit more of how JIM-Tcl works so they can do
something useful, and (2) those that want to add a new command to
OpenOCD.

@section Tcl Rule #1
There is a famous joke, it goes like this:
@enumerate
@item Rule #1: The wife is always correct
@item Rule #2: If you think otherwise, See Rule #1
@end enumerate

The Tcl equal is this:

@enumerate
@item Rule #1: Everything is a string
@item Rule #2: If you think otherwise, See Rule #1
@end enumerate

As in the famous joke, the consequences of Rule #1 are profound. Once
you understand Rule #1, you will understand Tcl.

@section Tcl Rule #1b
There is a second pair of rules.
@enumerate
@item Rule #1: Control flow does not exist. Only commands
@* For example: the classic FOR loop or IF statement is not a control
flow item, they are commands, there is no such thing as control flow
in Tcl.
@item Rule #2: If you think otherwise, See Rule #1
@* Actually what happens is this: There are commands that by
convention, act like control flow key words in other languages. One of
those commands is the word ``for'', another command is ``if''.
@end enumerate

@section Per Rule #1 - All Results are strings
Every Tcl command results in a string. The word ``result'' is used
deliberatly. No result is just an empty string. Remember: @i{Rule #1 -
Everything is a string}

@section Tcl Quoting Operators
In life of a Tcl script, there are two important periods of time, the
difference is subtle.
@enumerate
@item Parse Time
@item Evaluation Time
@end enumerate

The two key items here are how ``quoted things'' work in Tcl. Tcl has
three primary quoting constructs, the [square-brackets] the
@{curly-braces@} and ``double-quotes''

By now you should know $VARIABLES always start with a $DOLLAR
sign. BTW: To set a variable, you actually use the command ``set'', as
in ``set VARNAME VALUE'' much like the ancient BASIC langauge ``let x
= 1'' statement, but without the equal sign.

@itemize @bullet
@item @b{[square-brackets]}
@* @b{[square-brackets]} are command substitutions. It operates much
like Unix Shell `back-ticks`. The result of a [square-bracket]
operation is exactly 1 string. @i{Remember Rule #1 - Everything is a
string}. These two statements are roughly identical:
@example
    # bash example
    X=`date`
    echo "The Date is: $X"
    # Tcl example
    set X [date]
    puts "The Date is: $X"
@end example
@item @b{``double-quoted-things''}
@* @b{``double-quoted-things''} are just simply quoted
text. $VARIABLES and [square-brackets] are expanded in place - the
result however is exactly 1 string. @i{Remember Rule #1 - Everything
is a string}
@example
    set x "Dinner"
    puts "It is now \"[date]\", $x is in 1 hour"
@end example
@item @b{@{Curly-Braces@}}
@*@b{@{Curly-Braces@}} are magic: $VARIABLES and [square-brackets] are
parsed, but are NOT expanded or executed. @{Curly-Braces@} are like
'single-quote' operators in BASH shell scripts, with the added
feature: @{curly-braces@} can be nested, single quotes can not.  @{@{@{this is
nested 3 times@}@}@} NOTE: [date] is a bad example;
at this writing, Jim/OpenOCD does not have a date command.
@end itemize

@section Consequences of Rule 1/2/3/4

The consequences of Rule 1 are profound.

@subsection Tokenisation & Execution.

Of course, whitespace, blank lines and #comment lines are handled in
the normal way.

As a script is parsed, each (multi) line in the script file is
tokenised and according to the quoting rules. After tokenisation, that
line is immedatly executed.

Multi line statements end with one or more ``still-open''
@{curly-braces@} which - eventually - closes a few lines later.

@subsection Command Execution

Remember earlier: There are no ``control flow''
statements in Tcl. Instead there are COMMANDS that simply act like
control flow operators.

Commands are executed like this:

@enumerate
@item Parse the next line into (argc) and (argv[]).
@item Look up (argv[0]) in a table and call its function.
@item Repeat until End Of File.
@end enumerate

It sort of works like this:
@example
    for(;;)@{
        ReadAndParse( &argc, &argv );

        cmdPtr = LookupCommand( argv[0] );

        (*cmdPtr->Execute)( argc, argv );
    @}
@end example

When the command ``proc'' is parsed (which creates a procedure
function) it gets 3 parameters on the command line. @b{1} the name of
the proc (function), @b{2} the list of parameters, and @b{3} the body
of the function. Not the choice of words: LIST and BODY. The PROC
command stores these items in a table somewhere so it can be found by
``LookupCommand()''

@subsection The FOR command

The most interesting command to look at is the FOR command.  In Tcl,
the FOR command is normally implemented in C. Remember, FOR is a
command just like any other command.

When the ascii text containing the FOR command is parsed, the parser
produces 5 parameter strings, @i{(If in doubt: Refer to Rule #1)} they
are:

@enumerate 0
@item The ascii text 'for'
@item The start text
@item The test expression
@item The next text
@item The body text
@end enumerate

Sort of reminds you of ``main( int argc, char **argv )'' does it not?
Remember @i{Rule #1 - Everything is a string.} The key point is this:
Often many of those parameters are in @{curly-braces@} - thus the
variables inside are not expanded or replaced until later.

Remember that every Tcl command looks like the classic ``main( argc,
argv )'' function in C. In JimTCL - they actually look like this:

@example
int
MyCommand( Jim_Interp *interp,
           int *argc,
           Jim_Obj * const *argvs );
@end example

Real Tcl is nearly identical. Although the newer versions have
introduced a byte-code parser and intepreter, but at the core, it
still operates in the same basic way.

@subsection FOR command implementation

To understand Tcl it is perhaps most helpful to see the FOR
command. Remember, it is a COMMAND not a control flow structure.

In Tcl there are two underlying C helper functions.

Remember Rule #1 - You are a string.

The @b{first} helper parses and executes commands found in an ascii
string. Commands can be seperated by semicolons, or newlines. While
parsing, variables are expanded via the quoting rules.

The @b{second} helper evaluates an ascii string as a numerical
expression and returns a value.

Here is an example of how the @b{FOR} command could be
implemented. The pseudo code below does not show error handling.
@example
void Execute_AsciiString( void *interp, const char *string );

int Evaluate_AsciiExpression( void *interp, const char *string );

int
MyForCommand( void *interp,
              int argc,
              char **argv )
@{
   if( argc != 5 )@{
       SetResult( interp, "WRONG number of parameters");
       return ERROR;
   @}

   // argv[0] = the ascii string just like C

   // Execute the start statement.
   Execute_AsciiString( interp, argv[1] );

   // Top of loop test
   for(;;)@{
        i = Evaluate_AsciiExpression(interp, argv[2]);
        if( i == 0 )
            break;

        // Execute the body
        Execute_AsciiString( interp, argv[3] );

        // Execute the LOOP  part
        Execute_AsciiString( interp, argv[4] );
    @}

    // Return no error
    SetResult( interp, "" );
    return SUCCESS;
@}
@end example

Every other command IF, WHILE, FORMAT, PUTS, EXPR, everything works
in the same basic way.

@section OpenOCD Tcl Usage

@subsection source and find commands
@b{Where:} In many configuration files
@* Example: @b{ source [find FILENAME] }
@*Remember the parsing rules
@enumerate
@item The FIND command is in square brackets.
@* The FIND command is executed with the parameter FILENAME. It should
find the full path to the named file. The RESULT is a string, which is
substituted on the orginal command line.
@item The command source is executed with the resulting filename.
@* SOURCE reads a file and executes as a script.
@end enumerate
@subsection format command
@b{Where:} Generally occurs in numerous places.
@* Tcl has no command like @b{printf()}, instead it has @b{format}, which is really more like
@b{sprintf()}.
@b{Example}
@example
    set x 6
    set y 7
    puts [format "The answer: %d" [expr $x * $y]]
@end example
@enumerate
@item The SET command creates 2 variables, X and Y.
@item The double [nested] EXPR command performs math
@* The EXPR command produces numerical result as a string.
@* Refer to Rule #1
@item The format command is executed, producing a single string
@* Refer to Rule #1.
@item The PUTS command outputs the text.
@end enumerate
@subsection Body or Inlined Text
@b{Where:} Various TARGET scripts.
@example
#1 Good
   proc someproc @{@} @{
       ... multiple lines of stuff ...
   @}
   $_TARGETNAME configure -event FOO  someproc
#2 Good - no variables
   $_TARGETNAME confgure -event foo "this ; that;"
#3 Good Curly Braces
   $_TARGETNAME configure -event FOO @{
        puts "Time: [date]"
   @}
#4 DANGER DANGER DANGER
   $_TARGETNAME configure -event foo "puts \"Time: [date]\""
@end example
@enumerate
@item The $_TARGETNAME is an OpenOCD variable convention.
@*@b{$_TARGETNAME} represents the last target created, the value changes
each time a new target is created. Remember the parsing rules. When
the ascii text is parsed, the @b{$_TARGETNAME} becomes a simple string,
the name of the target which happens to be a TARGET (object)
command.
@item The 2nd parameter to the @option{-event} parameter is a TCBODY
@*There are 4 examples:
@enumerate
@item The TCLBODY is a simple string that happens to be a proc name
@item The TCLBODY is several simple commands seperated by semicolons
@item The TCLBODY is a multi-line @{curly-brace@} quoted string
@item The TCLBODY is a string with variables that get expanded.
@end enumerate

In the end, when the target event FOO occurs the TCLBODY is
evaluated. Method @b{#1} and @b{#2} are functionally identical.  For
Method @b{#3} and @b{#4} it is more interesting. What is the TCLBODY?

Remember the parsing rules. In case #3, @{curly-braces@} mean the
$VARS and [square-brackets] are expanded later, when the EVENT occurs,
and the text is evaluated. In case #4, they are replaced before the
``Target Object Command'' is executed. This occurs at the same time
$_TARGETNAME is replaced. In case #4 the date will never
change. @{BTW: [date] is a bad example; at this writing,
Jim/OpenOCD does not have a date command@}
@end enumerate
@subsection Global Variables
@b{Where:} You might discover this when writing your own procs @* In
simple terms: Inside a PROC, if you need to access a global variable
you must say so. See also ``upvar''. Example:
@example
proc myproc @{ @} @{
     set y 0  #Local variable Y
     global x #Global variable X
     puts [format "X=%d, Y=%d" $x $y]
@}
@end example
@section Other Tcl Hacks
@b{Dynamic variable creation}
@example
# Dynamically create a bunch of variables.
for @{ set x 0  @} @{ $x < 32 @} @{ set x [expr $x + 1]@} @{
    # Create var name
    set vn [format "BIT%d" $x]
    # Make it a global
    global $vn
    # Set it.
    set $vn   [expr (1 << $x)]
@}
@end example
@b{Dynamic proc/command creation}
@example
# One "X" function - 5 uart functions.
foreach who @{A B C D E@}
   proc [format "show_uart%c" $who] @{ @} "show_UARTx $who"
@}
@end example

@include fdl.texi

@node OpenOCD Concept Index
@comment DO NOT use the plain word ``Index'', reason: CYGWIN filename
@comment case issue with ``Index.html'' and ``index.html''
@comment Occurs when creating ``--html --no-split'' output
@comment This fix is based on: http://sourceware.org/ml/binutils/2006-05/msg00215.html
@unnumbered OpenOCD Concept Index

@printindex cp

@node Command and Driver Index
@unnumbered Command and Driver Index
@printindex fn

@bye