HP3000-L Archives

February 1999, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"Michael P. Smith" <[log in to unmask]>
Reply To:
Michael P. Smith
Date:
Fri, 12 Feb 1999 16:56:21 -0600
Content-Type:
text/plain
Parts/Attachments:
text/plain (51 lines)
Stan Sieler wrote in message ...
>Michael Smith (aka [log in to unmask], so he doesn't get this reply directly) writes:
>
>> happens when the program in question is not the one being run but is
>> actually being CREATEd by the program we are running?
>...
>> I think that the processes being CREATEd will use it's own internal
limits
>> (specified at :LINK time) or if none are specified it will use the system
>> limits (configured in :SYSGEN).  But if that is true, it would mean in
order
>> to change the limits on a program meant to run as a child, you would have
to
>> do an :LINKEDIT ALTPROG.  This would also mean that you couldn't have
>> different versions of the program running with different limits.  It just
>> doesn't sound right, but maybe that's what is happening.
>
>It's mostly right...you overlooked that the CREATEPROCESS intrinsic
>allows you to specify a value for NMSTACK and/or NMHEAP.  If you do that,
>it overrides the values (if any) in the program file and the system
>default values.
>
>Note:
>   IIRC, the CREATEPROCESS documentation fails to note that the
>   values for items 26 and 27 (NMSTACK and NMHEAP) are the *adddresses*
>   of the values, not the actual values.  And, for CM calls, they're
>   "word" addresses, not byte addresses.
>
>--
>Stan Sieler                                          [log in to unmask]
>                                     http://www.allegro.com/sieler.html
>
Thanks Stan.

It turns out that the user is not doing a CREATE or CREATEPROCESS but is
actualy doing a FORK().  It gets better.  If the program doing the FORK() is
linked or is ran with the NMHEAP option, the FORK() fails.  To top it off,
the newly created process, inherits his NMSTACK and NMHEAP (since this is
FORK()) and therefore will always have the system default.  Sounds like an
SR to me.

 ---------------------------------------------------------------
| Michael P. Smith                                              |
| Sr. Systems Programmer                                        |
| The Hertz Corporation    [log in to unmask] (work)          |
| Oklahoma City, OK        [log in to unmask] (personal)      |
 ---------------------------------------------------------------
The thoughts, views and expressions contained in this message are those
of Michael P. Smith, and do not necessarily reflect those of the Hertz
Corporation.  To reply via email, remove the .xx from email address.

ATOM RSS1 RSS2