hal - Introduction to the HAL API
HAL stands for Hardware Abstraction Layer, and is used by LinuxCNC to transfer realtime data to and from I/O devices and other low-level modules.
hal.h defines the API and data structures used by the HAL. This file is included in both realtime and non-realtime HAL components. HAL uses the RTPAI real time interface, and the #define symbols RTAPI and ULAPI are used to distinguish between realtime and non-realtime code. The API defined in this file is implemented in hal_lib.c and can be compiled for linking to either realtime or user space HAL components.
The HAL is a very modular approach to the low level parts of a motion control system. The goal of the HAL is to allow a systems integrator to connect a group of software components together to meet whatever I/O requirements he (or she) needs. This includes realtime and non-realtime I/O, as well as basic motor control up to and including a PID position loop. What these functions have in common is that they all process signals. In general, a signal is a data item that is updated at regular intervals. For example, a PID loop gets position command and feedback signals, and produces a velocity command signal.
HAL is based on the approach used to design electronic circuits. In electronics, off-the-shelf components like integrated circuits are placed on a circuit board and their pins are interconnected to build whatever overall function is needed. The individual components may be as simple as an op-amp, or as complex as a digital signal processor. Each component can be individually tested, to make sure it works as designed. After the components are placed in a larger circuit, the signals connecting them can still be monitored for testing and troubleshooting.
Like electronic components, HAL components have pins, and the pins can be interconnected by signals.
In the HAL, a signal contains the actual data value that passes from one pin to another. When a signal is created, space is allocated for the data value. A pin on the other hand, is a pointer, not a data value. When a pin is connected to a signal, the pin’s pointer is set to point at the signal’s data value. This allows the component to access the signal with very little run-time overhead. (If a pin is not linked to any signal, the pointer points to a dummy location, so the realtime code doesn’t have to deal with null pointers or treat unlinked variables as a special case in any way.)
There are three approaches to writing a HAL component. Those that do not require hard realtime performance can be written as a single user mode process. Components that need hard realtime performance but have simple configuration and init requirements can be done as a single kernel module, using either pre-defined init info, or insmod-time parameters. Finally, complex components may use both a kernel module for the realtime part, and a user space process to handle ini file access, user interface (possibly including GUI features), and other details.
HAL uses the RTAPI/ULAPI interface. If RTAPI is #defined hal_lib.c would generate a kernel module hal_lib.o that is insmoded and provides the functions for all kernel module based components. The same source file compiled with the ULAPI #define would make a user space hal_lib.o that is staticlly linked to user space code to make user space executables. The variable lists and link information are stored in a block of shared memory and protected with mutexes, so that kernel modules and any of several user mode programs can access the data.
For an explanation of realtime considerations, see intro(3rtapi).
Except as noted in specific manual pages, HAL returns negative errno values for errors, and nonnegative values for success.