Ring 3 Traps and the OS/2 Kernel Debugger (101462)
The information in this article applies to:
- Microsoft LAN Manager 2.2
This article was previously published under Q101462 SUMMARY
When an application running at ring 3 in OS/2 experiences a General
Protection Fault, Stack Fault, or any other unexpected exception, it
displays a short message, a trap value, and various register
information--all of which is referred to as a Trap Screen. The OS/2
kernel debugger is often used to address these problems but by default
the OS/2 kernel debugger does not activate on ring 3 traps. Therefore
the default settings of the OS/2 kernel debugger are useless in
analyzing ring 3 exceptions.
MORE INFORMATION
In order to get the OS/2 kernel debugger to activate for ring 3
exceptions you must specifically hook the interrupt vector using the
debugger's VSP command. For instance, if you are experiencing TRAP
000D (general protection fault), in your ring 3 application you will
need to enter the VSPD command on the debug terminal to hook the 0xD
protect mode interrupt. For a stack fault, TRAP 000C, enter VSPC to
hook the 0xC protect mode interrupt.
If you do not hook these interrupts the trap screen appears on the
server as if the debug kernel were not installed and all access to
code and data segments involved in the trap is lost.
REFERENCES
For more information on the VS(P) command and other debugger commands,
see the "OS/2 System Debugger Command Reference."
| Modification Type: | Major | Last Reviewed: | 10/23/2003 |
|---|
| Keywords: | KB101462 |
|---|
|