-
Notifications
You must be signed in to change notification settings - Fork 72
Troubleshooting
The following are known issues and their solutions or workarounds.
You may find some of the following commands useful during your debugging process:
Enumerate USB devices with device details
Linux:
lsusb -v
macOS:
ioreg -lfxp IOUSB
Monitor udev
rules on Linux
udevadm monitor
On startup hid4java
will search the classpath looking for a library that matches the machine OS and architecture (e.g. Windows running on AMD64). It uses the JNA naming conventions to do this and will report the expected path if it fails. Supported platforms include:
-
darwin-x86-64
- macOS 64-bit -
darwin-aarch64
- macOS ARM64 -
linux-aarch64
- Linux ARMv8 64-bit only -
linux-amd64
- Linux AMD 64-bit only -
linux-arm
- Linux ARMv7 hard float 32-bit only -
linux-armel
- Linux ARMv5 soft float 32-bit -
linux-riscv64
- Linux RISC-V 64-bit -
linux-x86-64
- Linux x86 64-bit only -
linux-x86
- Linux x86 32-bit only -
win32-x86
- Windows 32-bit only -
win32-x86-64
- Windows 64-bit only -
win32-aarch64
- Windows 64-bit ARM64
You can add your own entry under src/main/resources
and it should get picked up. Ideally you should raise an issue on the hid4java
repo so that the proper library can be put into the project so that others can avoid this problem.
If you would like to cross-compile hidapi
for your platform, please refer to these step-by-step instructions.
Check that the usage page is not 0x06
which is reserved for keyboards and mice.
Windows opens these devices for its exclusive use and thus hid4java
cannot establish its own connection to them.
You will most likely need to use the lower level usb4java
library for this as it is based on the libusb 1.0
native libraries. For more details on using libusb, particularly on Windows, have a read of this section of their documentation.
This is the default behaviour and has been maintained for backward compatibility (0.6.0
and below). From 0.7.0
onwards the HidSpecification
allows for a manual start (which was implied but not the actual behaviour) so that any attached devices can announce their presence immediately.
// Configure to use custom specification
HidServicesSpecification hidServicesSpecification = new HidServicesSpecification();
// Use manual start feature to get immediate attach events
hidServicesSpecification.setAutoStart(false);
// Get HID services
HidServices hidServices = HidManager.getHidServices(hidServicesSpecification);
hidServices.addHidServicesListener(this);
// Manually start the services to get attachment event
hidServices.start();
In order to detect if a device is attached or not hid4java
must periodically scan the attached devices list and compare them to its internal manifest. By default this is done at a fixed interval of 500ms. However some devices that take time to process data can become overloaded and so need "quiet time" to respond to a write request. To set this use
-
HidSpecification.scanMode
- the mode (no scanning, fixed interval, fixed interval with pause after write) -
HidSpecification.scanInterval
- the time between enumeration scans when free running -
HidSpecification.pauseInterval
- the time to pause until returning to free running until the next write occurs
By choosing an appropriate configuration from the above settings you should be able to reduce the enumeration load on your device.
// Configure to use custom specification
HidServicesSpecification hidServicesSpecification = new HidServicesSpecification();
// Use the fixed interval with pause after write to allow device to process data
// without being interrupted by enumeration requests
// (the default pause of 5000ms is sufficient for device and UX)
hidServicesSpecification.setScanMode(ScanMode.SCAN_AT_FIXED_INTERVAL_WITH_PAUSE_AFTER_WRITE);
This type of error is thrown by the Java Virtual Machine when a low level device library causes a problem.
One example of this is when JNA interacts with the hidapi
native library. If, for example, the getFieldOrder
list is given in the wrong order then the offsets used to find the exposed functions in the native library are calculated incorrectly and the JVM crashes. (In this case hid4java
uses the field list from Class.getFields() to ensure the correct order.
Another cause is if a Structure
has not been initialised and is being dereferenced, perhaps in a toString()
method.
Outside of the hid4java
operations, your application may be introducing USB conflicts - perhaps between hidapi
and libusb
. One approach is to choose which library will interact with which devices and then ensure that minimal overlap occurs. Perhaps hid4java
is scanning for devices (for attach/detach events) which could be handled by the other library. You can use ScanMode.NO_SCAN
to stop those events.
Further information on debugging low level messages from the JVM can be found in this StackOverflow answer.
You have probably terminated the JVM using a kill -9
rather than a clean shutdown. This will have left the HidApi
process lock on the DLL still in force and Windows will continuously check to see if it can share it with a new instance.
Just detach and re-attach the device to clear it.
Another explanation is that another process has grabbed the USB HID and has not released it. Check your task manager.
Finally, it may be that another process has gained an exclusive lock on the hidapi
native library (perhaps directly or via another JNA process) and it may require a reboot of Windows to ultimately clear it.
This was a device enumeration bug in early versions of hid4java
. Use version 0.3.1 or higher.
If you're seeing this then you should update hid4java
to the latest version.
This is a known issue with some mouse devices (e.g. Scorpion Marvo series) on Windows 10. The calling hidServices.scan()
method freezes the mouse pointer and there is currently no workaround (all scan modes give the same result). Standard USB mice do not have this issue.
Thanks to Brezgil for this catch.
In general there are two main types of device nodes in Linux: device drivers and raw device nodes. The first kind, such as a USB mass storage device would be mapped as /dev/sdb
. The second kind are general purpose raw USB devices. You will usually find references to hidraw devices under /dev/hidrawX
where X
is a number generated by udev
which is the userspace device manager for the Linux kernel (#include <libudev.h
in C). These raw devices are mapped as /dev/bus/usb/BBB/DDD
where BBB
is the bus number and DDD
is the device number. Using this approach bypasses the kernel driver.
Setting your own rules
Since a userspace program is acting like a device driver it has obvious security concerns and thus permissions need to be set correctly. By default, all raw USB device programs have 0664
permissions and need to have an entry under /etc/udev/rules.d
ending in .rules
. For example, /etc/udev/rules.d/99-myhid.rules
.
Typically when creating your .rules
file you will need the vendorId
and productId
. These can be found using lsusb
and examining the output for ID xxxx:yyyy
. xxxx
maps to vendorId
, yyyy
maps to productId
.
See this StackExchange answer on udev
for more details on setting rules. You may also find this introduction helpful.
Mapping /dev/hidraw
to /dev/bus/usb/BBB/DDD
Sometimes it is useful to determine which device under /dev/hidrawX
is mapped to which device in /dev/bus/usb/BBB/DDD
. This is non-trivial. You can start by finding BBB
and DDD
using lsusb
. The mapping to X
is provided through libudev
directly and will require compiling a utility app.
See USB device issues in Linux From Scratch for more information.
Different flavours of Linux require different settings:
Out of the box Ubuntu classifies HID devices as belonging to root. You can override this rule by creating your own under /etc/udev/rules.d
:
sudo gedit /etc/udev/rules.d/99-myhid.rules
Make the content of this file as below (using your own discovered hex values for idVendor
and idProduct
in lowercase with no 0x
prefix):
# My HID device
KERNEL=="hidraw*", ATTRS{idVendor}=="1234", ATTRS{idProduct}=="abcd", MODE="0666", GROUP="plugdev", TAG+="uaccess", TAG+="udev-acl"
(The TAG entries are there for compatibility with systemd
). You may find that you need to finalise MODE
with MODE:="0666"
as seen in issue #106 on Mint Linux (a Debian/Ubuntu variant).
Save and exit from root, then unplug and replug your device. The rules should take effect immediately.
If they're still not running it may be that you're not a member of the plugdev
group. You can fix this as follows (assuming that plugdev
is not present on your system):
sudo addgroup plugdev
sudo addgroup yourusername plugdev
Edit the USB udev rules /etc/udev/rules.d
as follows:
MODE="0666", GROUP="dialout"
Running on ARM machines you may encounter problems due to a missing library. This is just a naming issue for the udev
library and can be resolved using the following command (or equivalent for your system):
sudo ln -sf /lib/arm-linux-gnueabihf/libudev.so.1 /lib/arm-linux-gnueabihf/libudev.so.0
Thanks to @MaxRoma for that one!
This appears to be an issue with fwupd
on Ubuntu 16.0. As it is fixed in Ubuntu 18.0+ it is probably worth upgrading to that version.
However, if you cannot upgrade then a workaround is available in issue #97.
Thanks to @Laivindur for their detailed investigation.
The hidapi
Linux support library is available in two variants: libusb
and hidraw
. In general the hidraw
variant is the most flexible and it allows Bluetooth interfaces to be addressed. However, you can force the use of the older libusb
as follows (the default is hidraw
):
HidApi.useLibUsbVariant = true
Setting this parameter will enable libusb
libraries (which have slightly different behaviour) when executing on a Linux platform.