| Updated: October 28, 2024 |
The startup code configures the hardware, detects system resources, and loads and starts the OS kernel.
In order to keep the QNX Neutrino RTOS as configurable as possible, the startup code can program components such as the base timers, interrupt controllers, and cache controllers, as well as provide kernel callouts (see the Kernel Callouts chapter).
The startup code's main tasks are, in sequence:
If the IFS isn't in its final location in RAM when the startup code begins execution, the startup code copies it there.
The startup code itself can never be compressed. However, if the rest of the OS image is compressed, the startup code is responsible for decompressing it to its final location.
The startup code does the minimum needed to detect the hardware configuration, then completes any further configuration as required. Sometimes, all that's required is setting up the memory management unit (MMU), the timer hardware, and the interrupt controllers (see “Hardware initialization”).
Depending on the system, its software configuration can be either dynamically detected and set by the startup code, or pre-set (hardcoded).
When the startup code has initialized the system, it places the system information in the system page area, then transfers control to the OS kernel.
The system page is a dedicated area of memory. This area contains the data structures that the kernel examines to learn about the system configuration. This configuration information includes:
The system page also includes kernel callouts that provide board-specific code for:
For more information, see the System Page chapter.
The startup code's final task is to load the QNX Neutrino kernel and process manager (procnto*), which:
When this phase is complete, the system is fully operational. The OS kernel can proceed to load and run user processes, as specified in the build scripts (see “Scripts” in the OS Image Buildfiles chapter).