- Support for C/C++ applications
- Reliability, availability, and serviceability enhancements
- Performance improvements
Support for C/C++ applications
Support is added for 64-bit real addressing, which improves the performance and response time of applications that have very large memory and DASD storage demands, use Data in Memory, or need to access very large databases. New functions are added to the ptrace callable service (BPX1PTR) to support callers using the new general-purpose "high" registers.
Reliability, availability, and serviceability enhancements
- UNIX System Services Parmlib Limits Checking
This support provides enhancements for monitoring and managing UNIX System Services parmlib limits, including:
- Console messages that indicate the status of USS parmlib limits, allowing an installation to react quickly when limits are reaching critical levels.
- A new BPXPRMxx parmlib statement, LIMMSG(NONE|SYSTEM|ALL), which controls the displaying of these console message. A new SETOMVS operand, PID=, which dynamically changes a parmlib limit for a process. .
- A new keyword, LIMITS, on the DISPLAY OMVS command, which displays information about USS parmlib limits and current system usage. With the PID= keyword, LIMITS displays information for an individual process. With the RESET keyword, LIMITS resets the high-water marks for system limits to 0.
- Sysplex CDS Repair Tool
Corruption of the Couple Data Set (CDS) can prevent the USS file system from performing key sysplex operations. The Sysplex CDS Repair Tool, used under the direction of an IBM service representative, makes it possible to correct or isolate the scope of a defect in the CDS, reducing the need for a sysplex-wide IPL. The MODIFY operator command is enhanced to support these shared HFS diagnostic and repair functions.
- Support for MVS Dump Debugging
Post-mortem analysis is now possible on MVS dumps. A new dbx command line option, -C, puts dbx in full source-level debug mode.
Fast pthread quiesce and get
A new pthread_quiesce interface, BPX1PQG, is introduced, which freezes or unfreezes a set of threads and returns state information for them, without requiring that signals be sent to the target threads. This improves the performance and reliability of applications, such as JAVA, that use pthread_quiesce or signalling to stop threads.