System Page

The system page is a data structure with constants, references to other data structures, and a union shared by the different processor architectures that QNX Neutrino RTOS supports.

The system page contains essential information about the system, including the system page's size, information about the hardware platform (including IRQs and the CPU), the cache size, and the location of kernel callouts in memory.

One of the startup program's main tasks is to initialize the system page with the information that the OS needs to run on the current hardware (see Startup tasks in the Startup Programs chapter). After the startup program initializes the system page, it is populated with data that describes the current system and OS and applications can retrieve the information that they need.

System page location

The standard location for header files, including the syspage.h header file that defines the system page structure, is on your host system in ${QNX_TARGET}/usr/include, where ${QNX_TARGET} is the directory that holds the target-related components. For example, if the QNX SDP base directory is qnx700, the syspage.h file for 64-bit x86 targets is at the following location:


An SDP product patch or OS update can replace these header files (and other components such as binaries and libraries) with newer versions.
However, because QNX regularly updates header files between patches and updates, a specific BSP component often requires a newer version of a header file. When an updated header file is required, it is included with the BSP and the version in the SDP exists independently. When you build the BSP, the build process automatically uses the version shipped with the BSP. Any other coding operations you perform outside the context of the specific BSP use the SDP version of the header file.

System page contents

The syspage.h header file . Although the contents of the system page varies, it always has the following members:

struct syspage_entry {
	_Uint16t		size;
	_Uint16t		total_size;
	_Uint16t		type;
	_Uint16t		num_cpu;
	syspage_entry_info	system_private;
	syspage_entry_info	DEPRECATED_SECTION_NAME(meminfo);
	syspage_entry_info	hwinfo;
	syspage_entry_info	qtime;
	syspage_entry_info	callout;
	syspage_entry_info	callin;
	syspage_entry_info	typed_strings;
	syspage_entry_info	strings;
	syspage_entry_info	smp;
	syspage_entry_info	pminfo;
	_Uint32t		spare0[1];
	union {
		struct {
		_Uint32t			__x86_spacer;
		struct x86_syspage_entry 	x86;
		struct {
		_Uint32t			__arm_spacer;
		struct arm_syspage_entry	arm;
		struct x86_64_syspage_entry	x86_64;
		struct aarch64_syspage_entry	aarch64;
		_Uint64t			filler[20];
	} un;

To see the exact definition of the system page for your system, you can look at syspage.h.

Note: The startup library code initializes some of the system page fields. Depending on the amount of customization you require, you may need to initialize some fields as well.

The list below describes the members of the syspage_entry data structure:

The size of the system page entry. The startup library sets this member automatically.

The size of the system page entry plus the referenced substructures; effectively the size of the entire system-page database.

The startup library sets this member automatically, but other library calls may adjust (grow) it later as required.


This member is used to indicate the CPU family for determining which union member in the un element to use. It can be one of SYSPAGE_X86 or SYSPAGE_ARM, for the supported 32-bit CPU architectures; or one of SYSPAGE_X86_64 or SYSPAGE_AARCH64, for the supported 64-bit architectures.

The library sets this member automatically.

The number of CPUs present on the system. The startup library initializes it to the default value 1. If additional CPUs are detected during startup, the startup program should call init_smp() to adjust the value of this member to the correct value.
See system_private.
See asinfo.
Deprecated; do not use.
See hwinfo.
See cpuinfo.
See cacheattr.
See qtime.
An system page area set aside for kernel callouts (see the Kernel Callouts chapter).
For internal use.

An area with several string entries; each entry consists of a four-byte number and a NULL-terminated string. The number corresponds to a specific constant from the system include file <confname.h> (see the C function confname()).

Usually, the init_*() library functions populate the typed strings literal pool. However, if you need to add something, you can call the add_typed_string() function from the startup library.


A literal pool used for non-typed strings. To use a non-typed string, specify the appropriate index into strings (e.g., cpuinfo's name member).

Usually, the init_*() library functions populate the literal pool. However, if you need to add something, you can call the add_string() function from the startup library.

See intrinfo.
See smp.
Not currently supported.
See un.


The _SP_NEW_ARRAY macro defines the data structures for several syspage_entry members, including asinfo, cpuinfo, cacheattr, intrinfo, and mdriver. It is a union of the old and the new syspage_array_info structures:

#define _SP_NEW_ARRAY(__sect)				\
	union {					\
		syspage_array_info	new_##__sect;	\
		syspage_array_info	__sect;	\


The _SP_OLD_ARRAY macro defines pre-64-bit data structures for the same syspage_entry members as are defined by _SP_NEW_ARRAY:

#define _SP_OLD_ARRAY(__sect)	syspage_entry_info	old_##__sect


The syspage_array_info data structure is used by _SP_NEW_ARRAY and _SP_OLD_ARRAY. It is defined as follows:

typedef struct {
	_Uint16t		entry_off;
	_Uint16t		entry_size;
	_Uint16t		element_size;
} syspage_array_info;

64-bit support

QNX SDP 7.0 introduces support for 64-bit processing on x86 and ARM platforms. This support introduces changes to the system page, including the addition of new definitions for system page types:

 *	System page types
enum {
	/* additional 32 bit architectures go here */

	SYSPAGE_64BIT = 0x100,
	/* additional 64 bit architectures go here */
Note: Some of the definitions are for architectures that are no longer supported (PPC, MIPS, SH).