🚀 go-pugleaf

    I am writing a utility in C to find all RMS locks on a specified file.
The scheme is:

(a) Call SYS$PARSE and SYS$SEARCH to find the 6-byte file ID (field NAM$W_FID)
    and device name (field NAM$T_DVI).  The file could be specified by a
    logical.

(b) Call SYS$GETDVI with item code DVI$_DEVLOCKNAM to find the 16-byte device
    lock name for the device name obtained in (a).

(c) Call SYS$ENQ to queue a lock on a resource whose name is concatenated as

        RMS$ + file ID + device lock name

    and receive a lock-id.

(d) Call SYS$GETLKI with item code LKI$_LOCKS, passing lock-id returned in
    (c), in hope to get a list of RMS locks associated with the resource name
    in (c).

(e) Print out PID, lock mode and queue for every lock in the list returned
    in (d).

    I have coded a test program.  My questions are:

1)  Can the above scheme work?  Is the basic idea correct?

2)  Apparently I have no access to RMS locks in normal situation.  What
    happen if I do (c) and (d) in EXEC mode (as same as RMS)?  If I queue
    a lock in EXEC mode with a valid resource name of a RMS file lock, can
    I block locks queued by RMS on the same file?  :-)

3)  If I shall change mode and call SYS$CMEXEC (or SYS$CMKRNL), what is the
    order of the argument list?  In the same order as in C or in a reversed
    order?

    I would like to have your experts' opinions before I go further, I am
a bit anxious about to shut down a busy system unwittingly.  Your help is
highly appreciated.  I thank Dan Wing, Earl D. Lakia and Martin P.J. Zinser
for their help on a previous request for lock utility.


-- Sanyee Tang
   BC Research
   Vancouver, Canada