Barbara Nitz wrote:
We finally IPL'd for the first time with all dirty getmain traps active, and it
was a desaster. The first to fold was Beta Systems Subsystem Initialization
(actually, BSA code). They get a pic4 attempting to store into PSA with the
guilty register bein x'FFFFFFFF'. What I don't know is what the problem
described in that Beta flash would look like.
We had identified two homegrown Assembler programs (one an SMF exit - fortunately SMF disables the exit after an abend!) beforehand, gotten rid of one and 'fixed' the other the wrong way.
Our big offender is actually BMC, with at least 5 problems in a lot of
different address spaces.
The FFFFFFFF probably came from the IeaInitRegsTask trap. It's pretty
scary when you come across code that uses a base register without
loading it first. (It's an easy mistake for a programmer to make.) The
scary part is, for authorized code, it's a potential integrity exposure!
IeaInitRegsTask and IeaInitArSrb do a good job in helping to uncover
these sorts of problems.
IgvInitGetmain doesn't "dirty" all getmained storage. It only "dirties"
storage not guaranteed to be zeros by the operating system. (See the
CHECKZERO=YES keyword on the GETMAIN macro.) Assuming zeroes, when the
operating system doesn't guaranteed zeros, is an exposure to abend or
worse--a potential system integrity issue if authorized code.
--
Edward E Jaffe
Phoenix Software International, Inc
5200 W Century Blvd, Suite 800
Los Angeles, CA 90045
310-338-0400 x318
[email protected]
http://www.phoenixsoftware.com/
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
