Problem beim Portzugriff unter NT (betrifft TC_832)

aus MSDN 6.0

DOCERR: Port I/O with inp() and outp() Fails on Windows NT
Last reviewed: January 19, 1996
Article ID: Q112298
The information in this article applies to:
The C Run-time (CRT), included with: Microsoft Visual C++ 32-bit Edition,
versions 1.0 and 2.0

Attempting to use the port I/O (input/output) functions such as inp() and
outp() from within an application for Windows NT running in user mode causes a
privileged instruction exception to occur. The sample code provided in the
OUTP.C online help sample for the outp() function and related functions can be
used to illustrate this behavior.

The documentation regarding the compatibility of the port I/O functions is
incomplete. Win32-based applications that call inp(), outp(), and so forth can
be successfully compiled and linked. However, these applications will generate
the privileged instruction exception because the port I/O functions cannot be
called from code running in user mode.

Do not call the following functions from within a Win32-based application
executing in user mode:


NOTE: This documentation error has been corrected for the Visual C++ 4.0 Books
Online. The descriptions of the various port I/O functions do not list "Win NT"
in the Compatibility section.

According to the documentation provided with Visual C++ 32-bit Edition, inp(),
outp(), and the other port I/O related functions are Win32 and Win32s
compatible. However, executing code that uses these functions causes a
privileged instruction exception on Win32 on Windows NT. The code that uses
inp(), outp(), and so forth will execute correctly on Win32s.

The inp(), outp(), and other I/O port related functions map to privileged
processor instructions. For example, on Intel processors, the inp() and outp()
functions end up calling the IN and OUT instructions. The privileged
instruction exception occurs when these instructions are executed because
typical Windows NT applications execute in a nonprivileged (user) mode. Only
code executing in kernel mode has the necessary rights to execute privileged
instructions. Kernel mode code is typically found in device drivers.

For more information on user mode and kernel mode in Windows NT, refer to the
Microsoft Press book "Inside Windows NT" by Helen Custer. For an example of a
kernel mode Windows NT device driver that allows user mode applications to
access hardware ports, refer to the GENPORT sample provided with the Microsoft
Windows NT Device Driver Kit (DDK).

The Online help for _inp, _inpw, _inpd, _outp, _outpw, and _outpd. The
Microsoft Press book "Inside Windows NT" by Helen Custer. The Microsoft Press
book "Microsoft s 80386/80486 Programming Guide" by Ross P. Nelson. The
"Kernel-Mode Device Driver Guide" provided with the Microsoft Windows NT DDK.

Hardware Abstraction Layer (HAL)

Windows NT is designed to run with a variety of CPUs and hardware platforms.
This places a heavy burden on the kernel to efficiently reconcile kernel
processing with the need to handle interrupts, cache management, memory
handling, register assignments, and other hardware-specific implementations. To
solve this portability issue, Windows NT isolates the kernel code from the
hardware by using an isolation layer referred to as the hardware abstraction
layer, or HAL. The HAL presents a "virtual machine" to the kernel, executive,
and device drivers. It maps this virtual machine onto the underlying hardware
in an efficient manner and eliminates the need for Windows NT to provide
hardware specific kernels, which improves overall efficiency. The HAL presents
a single mechanism for device driver and kernel components to use on any
hardware platform.

As an example, in Intel(r)-based systems, device drivers use the INP and OUTP
instructions to access the I/O port registers directly. Other CPU architectures
use different mechanisms for accessing device I/O registers. Windows NT
resolves this difference with macros (such as read_port_uchar and
write_port_uchar) that take the appropriate action for each supported CPU.