Improve errpt CORE_DUMP entries to refer back to executable image that was running

When using dbx to analyze a core file that was dumped, the executable binary that was running must be used in combination with the core file to identify the exact instruction that caused the core dump.  In the errpt CORE_DUMP entry there is the full pathname of the core file and entries named "FILE SYSTEM SERIAL NUMBER" and "INODE NUMBER".  Per the https://www.ibm.com/docs/en/aix/7.2?topic=overview-error-logging-tasks these are “Information on the current working directory of the process, such as FILE SYSTEM SERIAL NUMBER and INODE NUMBER when the process dumps the core.”  However, it makes much more sense if these values are references to the executable binary image that was running within the limits of how Unix executes a file.

When an executable is started the INODE on the FILE SYSTEM are linked to by the OS and that file mapped to memory (in read-only mode) and the execution of the binary started. If the directory entry to the executable file is removed or renamed, the link to the inode by the OS keeps the memory map valid. Consequently the FILE SYSTEM/INODE numbers are the only information that can be guaranteed to be valid after the executable starts. That is why they would be returned in the errpt in the CORE DUMP message, they are the only things that can be certain to be valid at the time of the issue that caused the core dump.

It should be noted that in AIX Core Dump Data Collection https://www.ibm.com/support/pages/aix-core-dump-data-collection this indicates that for an errpt CORE_DUMP entry the INODE NUMBER is that of the executable that was running. Specifically they take the INODE NUMBER and run find with it to get the full path of the executable.