mxt-app - command line utility for maXTouch devices
mxt-app [command] [options]...
mxt-app is a utility for managing Atmel maXTouch touch controllers and other devices that support Atmel Object Based Protocol.
If no command is not given, mxt-app will provide an interactive menu based interface.
The Atmel Object Based Protocol defines how device registers (normally accessed via I2C) are mapped to different functions within the devices. This interface organises the register map into separate objects each of which is given a T number. mxt-app can inspect and alter the object configuration, and view diagnostic data, while the device is running.
For a description of object protocol, see Atmel AT42QT1085 Object Protocol Guide, available from atmel.com.
The meaning of the configuration bytes within the objects may be found in the Protocol Guide documentation released with each device, and is only provided by Atmel under NDA.
-h [--help]
: Display a brief summary of available options and exit.
-i [--info]
: Print the ID information and object table.
-M [--messages] [*timeout*]
: Prints the messages until timeout seconds have passed. If no timeout
is provided, continue until user presses Ctrl-C. Zero timeout reads once.
Provide -F [--msg-filter]
option to filter by a specific object.
-F [--msg-filter] *TYPE*
: Filters messages by object TYPE.
--reset
: Reset device.
--calibrate
: Send calibrate command.
--backup[*=COMMAND*]
: Backup configuration to NVRAM where the optional argument, COMMAND, is the BACKUPNV command.
-g
: Write Golden Reference calibration to NVRAM.
--self-cap-tune-config
: Tune and calibrate the self capacitance settings and store them to the device configuration.
--self-cap-tune-nvram
: Tune and calibrate the self capacitance settings and store them to NVRAM without updating the Config Checksum.
--version
: print version of mxt-app.
--block-size *BLOCKSIZE*
: Sets the i2c block size.
--load *FILE*
: Upload config from FILE, write it to NVRAM, and reset device. The
configuration may be in .xcfg
or OBP_RAW
format.
--save *FILE*
: Save config to FILE in either OBP_RAW
or .xcfg
format.
--checksum *FILE*
: Read the contents of FILE and recalculate the configuration checksum.
-R [--read]
: Read data from the device.
-W [--write]
: Write data to the device.
-n [--count] *COUNT*
: read/write COUNT registers
-f [--format]
: format register output
-I [--instance] *INSTANCE*
: select object INSTANCE
-r [--register] *REGISTER*
: start at REGISTER (offset in object when used with TYPE)
-T [--type] *TYPE*
: select object TYPE
--zero
: zero all configuration settings
$ mxt-app -R -n7 -r0
82 19 11 AA 18 0E 16
$ mxt-app -R -T7
32 FF 05 43
$ mxt-app -W -T7 0000
$ mxt-app -R -T7 --format
GEN_POWERCONFIG_T7
00: 0x00 0 0000 0000
01: 0x00 0 0000 0000
02: 0x05 5 0000 0101
03: 0x43 67 0100 0011
mxt-app supports connection over TCP using a ASCII protocol which allows mxt-app to act as a bridge so that Atmel proprietary tools such as Object Server can access the device.
-C [--bridge-client] *HOST*
: Connect over TCP to HOST
-S [--bridge-server]
: Start TCP socket server
-p [--port] PORT
: TCP port (default 4000)
--bootloader-version
: Query and print ID and version of bootloader.
--flash *FIRMWARE*
: Flash FIRMWARE to device. The firmware file should be in .enc
format.
--reset-bootloader
: Reset device in bootloader mode. In bootloader mode the device will cease
normal operation until a firmware is sent. The I2C address or USB PID will
change. The only valid command in this mode is --flash
. A hard power
cycle will return the device to normal Object Protocol mode, unless the
firmware image is corrupted. This command is only provided for debugging
purposes: in most cases --flash
will manage the change to/from
bootloader mode before/after flash.
--firmware-version *VERSION*
: The .enc file format does not provide the firmware version in a form
available to mxt-app. If it is provided via this switch, mxt-app can check
firmware VERSION before and after flash. It will skip the flash process
if the firmware version is already correct. It will also check for a
successful flash on completion. The version must be provided in the format
1.0.AA
.
The Self Test T25 object runs self-test routines in the device to find faults in the sense lines and electrodes. The Self Test T25 object runs a series of test sequences.
-t [--test]
: Run all self tests.
-t*XX* [--test=*XX*]
: Run individual self test specified by the CMD hex value.
-t01
: run analog power test.
-t11
: run pin fault test.
-t12
: run pin fault 2 test.
-t13
: run AND gate test.
-t17
: run signal limit test.
-t20
: run gain test.
-t21
: run offset fault test.
Capture frames of diagnostic data. The default mode is to capture touch deltas. Self capacitance measurements are only available on some devices.
--debug-dump *FILE*
: The T37 Diagnostic Data object provides raw access to touch reference/delta
measurements from the touch screen. Diagnostic data is written to FILE in
CSV format. Format 0 is compatible with the Atmel Hawkeye utility.
--frames *N*
: Capture N frames of data.
--instance *N*
: Capture object instance N. Defaults to instance 0.
--format *N*
: Capture using Format 0 or 1.
Format 0 - Outputs all nodes in single line (X0Y0, X0Y1, ... X1Y0).
Format 1 - Outputs in (X) row and (Y) column format.
--references
: Capture references data.
--self-cap-signals
: Capture self cap signals.
--self-cap-deltas
: Capture self cap deltas.
--self-cap-refs
: Capture self cap references.
--active-stylus-deltas
: Capture active stylus deltas.
--active-stylus-refs
: Capture active stylus references.
--t68-file *FILE*
: Upload FILE to the device via the T68 Serial Data object.
--t68-datatype *DATATYPE*
: Set DATATYPE of the file. This will be automatically detected from the
file itself in most cases.
The broken line test scans the diagnostic data to identify touch sensor defects.
--broken-line
: Run the broken line detection algorithm on the device
--dualx
: Indicate if sensor X lines are double connected
--x-center-threshold N
: Set threshold for X lines in center of sensor to N percent
--x-border-threshold N
: Set threshold for X lines at edge of sensor to N percent
--y-center-threshold N
: Set threshold for Y lines in center of sensor to N percent
--y-border-threshold N
: Set threshold for Y lines at edge of sensor to N percent
--pattern *PATTERN*
: set sensor PATTERN material to ITO or Xsense
The sensor variant algorithm calculates all nodes of a channel to find the trend line of the reference. A comparison is done between the real and expected sensor reference.
--sensor-variant
: Perform the Sensor Variant algorithm
--dualx
: Indicate if sensor X lines are double connected
--fail-if-any
: Fail the Sensor Variant test on any defects
--max-defects N
: Maximum No. of continuous defects
--upper-limit N
: Upper limit for regression, in percentage %
--lower-limit N
: Lower limit for regression, in percentage %
--matrix-size N
: The allowed matrix size
By default mxt-app will scan available devices and connect to the first device it finds.
-q [--query]
: Scan for devices and output a list.
-d [--device] *DEVICESTRING*
: Connect to a particular device specified by DEVICESTRING which is given
in the same format as output by --query
.
There are three connection methods supported for hardware access:
This is used in conjunction with the Linux kernel driver. It accesses sysfs attributes under the directory
/sys/bus/i2c/drivers/dddddddd/b-00xx/
Where
d
: driver name - atmel_mxt_ts
, Atmel MXTXXXX
, etc
b : i2c adapter
xx : i2c address
A specific USB device can be specified by giving a device
option -d sysfs:PATH
as given by -q/--query
option
The sysfs attributes used under this directory are
mem_access
: Access to raw I2C address space.
debug_enable
: Output messages from the device to dmesg log as hexadecimal.
debug_v2_enable
, debug_msg
, debug_notify
: Optional improved binary interface to retrieve messages
They are provided when using the Atmel kernel driver from github, and may be supported by other devices.
Many maXTouch devices support a USB mode which reports touches via USB HID. In addition, evaluation boards may use a "bridge chip" which interfaces I2C to the same protocol.
USB mode will be built by autotools when libusb is available.
A specific USB device can be specified by giving a device
option -d usb:001-003
which corresponds to the bus and device numbers
given by the -q/--query
option and lsusb.
Devices can be accessed directly via the i2c-dev I2C debug interface by giving adapter and address on command line.
The i2c-dev interface is documented in the Linux kernel source, in Documentation/i2c/dev-interface
The I2C debug interface support must be enabled using the CONFIG_I2C_CHARDEV
kernel configuration option. It is enabled on a system if files /dev/i2c-*
are present.
To use i2c-dev, provide a device string such as -d i2c-dev:1-004a
.
Messages from the maXTouch devices are read by polling. If a kernel driver is also present on the system, reading messages on interrupt, then no messages will be received by the tool. A workaround is to set T18.COMMAND (byte 1) to 2 "Force the CHG line high (inactive)" so the kernel driver does not receive an interrupt.
There is no scanning support. This is because reading from every possible maXTouch address on every I2C bus might adversely affect some unrelated hardware that does not understand Object Protocol. You must manually identify the correct adapter and address by reference to the protocol guide or to the platform setup.
It is possible to use the --flash
command with a device already in bootloader
mode, by specifying the bootloader address.
The hidraw backend supports maXTouch devices which connect using USB or HID over I2C.
The hidraw interface is documented in the Linux kernel source, in Documentation/hid/hidraw.txt
The device must have /dev/hidraw raw HID device support enabled using the CONFIG_HIDRAW kernel configuration option.
To use hidraw, provide a device string such as -d hidraw:/dev/hidraw0
.
There is no scanning support.
Bootloading is not supported in this mode.
-v [--verbose] *LEVEL*
: set debug level. LEVEL is one of 0 (Silent), 1 (Warnings and Errors),
2 (Info - default), 3 (Debug), 4 (Verbose). Debug and Verbose are
only available if built in.
0 : Success
1 : Internal error/assert
2 : Input/output error
3 : Memory allocation failure
4 : Timeout
5 : Could not find a device or device went away
6 : Permission denied
7 : Operation not allowed for this device type
8 : Interrupt function call
9 : Object not available on device
10 : Received unexpected invalid message from message processor
11 : Self test invalid test command
12 : Self test AVdd Analog power is not present
13 : Self test Pin fault
14 : Self test AND Gate Fault
15 : Self test Signal limit fault
16 : Self test Gain error
17 : Information block checksum error
18 : Bootloader already unlocked
19 : Bootloader CRC failure (transmission failure)
20 : File format error
21 : Device firmware already required version
22 : Could not identify bootloader address
23 : Version on device did not match version given after bootloading operation
24 : Device did not reset
25 : Device in unexpected state
26 : Incorrect command line parameters or menu input given
27 : Bridge TCP protocol parse error
28 : Bridge connection error
29 : Serial data download failed
30 : No such file or directory
31 : Error processing self cap command
To download the source code using git:
git clone https://github.com/atmel-maxtouch/mxt-app.git
There are two build harnesses, for Android and autotools:
To download libusbdroid submodule:
git submodule init
git submodule update
Or to build without usb support:
MXTAPP_NO_USB_SUPPORT := true
To compile using Android NDK:
ndk-build
To enable debug:
ndk-build NDK_DEBUG=1
To disable usb support:
ndk-build MXTAPP_NO_USB_SUPPORT=true
To enable PIE support (for Android L):
ndk-build APP_PLATFORM=android-16
Binaries will be placed in libs/
The Android NDK is available from https://developer.android.com/tools/sdk/ndk/
adb push libs/armeabi/mxt-app /data/local/tmp/
adb shell /data/local/tmp/mxt-app [command]
If executable permissions have not been set, run:
adb shell chmod 777 /data/local/tmp/mxt-app
To compile using autotools:
./autogen.sh && make
To cross-compile:
./autogen.sh --host=arm-linux-gnueabi && make
or,
./autogen.sh --host=arm-linux-gnueabihf && make
To enable debug:
./autogen.sh --enable-debug
To enable generation of the man page using pandoc:
./autogen.sh --enable-man
To build the doxygen documentation (this requires doxygen and graphviz to be installed):
make doc
A version number is generated by git describe
during the build process and
is reported by --version
and to debug logs.
A typical version might be 1.15-29-g8321
which means, 29 commits after the
release tag 1.15, with a git SHA id beginning with 8321.
If the source is not checked out using git (for example by clicking on the github "Download ZIP" link), then the version from the file VERSION in the source archive is used.
The suffix -mod
is appended if there are uncommitted changes in the source
code.
If you see the warning
W: klogctl error 1 (Operation not permitted)
this indicates that mxt-app has been unable to retrieve messages from dmesg. Various features will not work properly. It may be possible to unrestrict dmesg by doing
# echo 0 > /proc/sys/kernel/dmesg_restrict