H-Series Application Migration Guide (H06.03+)

General Programming Issues

H-Series Application Migration Guide—429855-006

4-3

Misalignments in TNS/E Native Programs

On TNS/E systems, round-down never occurs in native mode, and no options to

change native mode program behavior are supported. Certain misalignments are

treated as fatal events, and for those, no tracing is necessary. Nonfatal misalignments

can also occur, and on both G-series and H-series systems, they can cause

compatibility traps. A program that spends excessive time in compatibility traps caused

by data misalignment can experience severe performance loss. This performance loss

is even greater on TNS/E systems that on TNS/R systems.

To ensure consistent program behavior when migrating TNS/R native programs to

TNS/E native mode and to avoid any performance loss, you should correct data

misalignments before migration. Use the available options to log the misalignment

events or cause the program to fail, then fix the program.

Detecting Data Misalignments

On H-series systems, you can use the EPTRACE facility for reporting misalignment

events and compatibility traps in TNS/E native processes without halting program

execution. EPTRACE is a fully supported version of the unsupported TNS/R XLTRACE

facility, and can be safely used in a production environment.

When run against a program that contains misaligned addresses, EPTRACE reports

the type of each trap and the code address of where each trap occurred. Execution

continues after the event. You can then use the enoft utility to map the code address

from EPTRACE to the source code location in the program.

EPTRACE runs only in the Guardian environment. See the EPTRACE Manual for more

information about using EPTRACE.

The following documents provide more information on detecting and correcting data

misalignments:

C/C++ Programmer’s Guide

pTAL Reference Manual

COBOL Manual for TNS/E Programs

Adjusting For Increased DCT Limits

The Destination Control Table (DCT) contains entries for logical device numbers and

named processes. The DCT limit refers to the maximum number of logical device

numbers and named processes that the operating system can accommodate. As of the

G06.23 RVU, the size of the DCT can optionally be increased from its previous limit of

32,767 (a logical device number can have at most 15 bits) to 65,376 (a logical device