>> To be fair to our boy Rob, I don't think he's actually advocating using two
>> LXs. He is just pointing out that the scheme you're proposing would *need*
>> two.
Exactly.
One could make a case for non-SysLX usage without a SysLX PC-cp to ATSET+ETCON
if the "client" address spaces were all capable of supervisor state anyhow.
For example, a multi-ASID product where certain requests must be squirted over
to one specific queue in the private area of one specific ASID.
Going back to a more all-encompassing server that caters for problem and
supervisor clients, you could justify the two LX design if the PC-cp did the
SAF checking for the server. If implemented this way, the non-SysLX PC
routine(s) cannot be executed due to XMS architecture reasons if the SAF check
in the PC-cp did not pass (and the ATSET and ETCON were not performed). If the
number of non-SysLX routines is non-trivial this could be beneficial to the
product to centralize the security checking.
Rob Scott
Lead Developer
Rocket Software
77 Fourth Avenue . Suite 100 . Waltham . MA 02451-1468 . USA
Tel: +1.781.684.2305
Email: [email protected]
Web: www.rocketsoftware.com
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf
Of Chris Craddock
Sent: 18 February 2013 19:34
To: [email protected]
Subject: Re: Load To Global with PC_cp
>>On Mon, Feb 18, 2013 at 10:32 AM, [email protected] <[email protected]>
wrote:
>>Rob Scott wrote
>>>(a) The first will have LXRES SYSTEM=YES for the PC-cp - this will be
>automatically available to ALL address spaces
>>>(b) The second will have LXRES SYSTEM=NO for the other PC-ss routines
that >your "selected" address spaces will use
>>Its interesting Rob recommended the reservation of two Linkage Indexs.
To be fair to our boy Rob, I don't think he's actually advocating using two
LXs. He is just pointing out that the scheme you're proposing would *need* two.
That's the main reason that almost nobody uses non-system LXs. You need a PC
service already connected to a system LX in order to connect the non-system LX
to each client address space. And since you would by definition already have a
system LX, the non-system LXs would be superfluous.
One could argue there are benefits in using non-system LXs but I don't know of
anybody that does that. Most long-lived vendor code just gives the server space
an AX of 1 and uses a single system LX and pc entry table. No muss, no fuss.
That's why IBM ended up having to invent that ASN and LX reuse feature...
(topic for another day)
--
This email might be from the
artist formerly known as CC
(or not) You be the judge.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send email to
[email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
