gtpm1m35TPF V4R1 Migration Guide: 3.1 to 4.1

System Error Support

Before this release, you could specify the content of dumps in two different ways:

Dumps were divided into three sections:

Beginning with this release, the TPF dumps are divided into two sections:

You can use the new optional PREFIX parameter on the SERRC macro to define multiple sets of system error numbers. The prefix character you specify will appear at the beginning of the system error number. All IBM system errors begin with the letter I. The letters W through Z are reserved for future use by IBM. User system errors will default to the letter U.

Use the new LIST parameter to specify additional main storage areas to be dumped. The CLIST parameter is no longer supported in the TPF 4.1 system. The SLIST parameter is still supported. See Controlling Dump Content for more information.

For operational (OPR) dumps, you only get the ECB virtual memory (EVM) unless you enter the ZASER command with the LONG parameter specified. For control (CTL) dumps, the SMDT replaces the selective dump address table.

In addition, in the TPF 4.1 system:

You can also start the dump data user exit to present ECB-controlled SERRC and SNAPC data. For SERRC dumps, a subset of the dump data will be presented to the user exit.

The new DATAX and NODATAX parameters for the ZASER command activate and deactivate this user exit. The dump data user exit cannot be controlled independently. The ZASER command with the DATAX parameter specified turns on the data exit for both SNAPC and SERRC dumps.

See TPF Main Supervisor Reference for more information about system error support.

Modifying Dump Tags

The format of the alter dump tag (ZADCA command) has changed and there is a validation option allowing you to verify up to 16 bytes of data that you want to modify. If the validation data does not match, an error message is issued and no modification is done.