On Mon, 16 Mar 2015 12:44:20 -0700, Charles Mills wrote:

>> if it were not designed to be invoked in an authorized environment, it 
>> should not be included in an APF authorized load library
>
Alas, that's an impossible goal, or at least impractical given the legacy 
milieu.
I suspect there'd be much breakage if even IEFBR14 were removed from
SYS1.LINKLIB and placed in an unauthorized library:

o Any CLIST containing CALL 'SYS1.LINKLIB(IEFBR14)' would fail.
o Any authorized program using IEFBR14 as a stub/default would fail.
(I doubt that IEFBR14 introduces any integrity exposure.)

>2. ..., and it is the responsibility of any authorized caller to be certain 
>(how?) that calling it authorized will not generate an integrity exposure.
>
When I whined about the "(how?)" in connection with SMP/E a few years ago,
before I knew even what little I now suspect about the nature of the weakness,
Walt replied with words similar to "reasonable caution".  I take that to mean
that whatever flaw, it's (perhaps) susceptible to malicious exploitation, but
highly unlikely to be triggered inadvertently.  It's not necessarily IEBCOPY, 
but
any program a programmer can name in a UTILITY entry.

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to