On Tue, 15 Dec 2009 08:59:41 +0800 David Stephens <[email protected]> wrote:
:>Thanks Rob, Chris and Binyamin for your comments. I've learnt a lot from :>them. :>A couple of questions for the panel: :>Chris, you said that: :>"Using AX=1 and SSAR to another address space is not very clearly documented by IBM because it is a violation of the MVS xmemory architecture". :>This comes as a surprise, as I've seen this used in many different programs. Setting AX=1 will certainly allow you to access any address space's memory, but unless you have another TCB needing to AXSET to something other than 1, I can't see how this may not work, or could cause difficulties. Can you give more details? :>You also said that using ASC mode doesn't remove the need for the target address space to be swapped in. Short of a paging error, what could go wrong here? :>Binyamin, you've suggested grabbing the LOCAL lock when the SSAR succeeds. Why do you suggest this? The basic issue is that AX=1 & SSAR gives you a hardware bind to the address space, but MVS is not "aware" of it. It is possible that MVS will choose to swap out what you are using as the secondary address space. MVS does not take kindly to a page fault against a swapped out address space. Taking the LOCAL lock of the secondary address space informs MVS that you have a bind and prevents the swap out. -- Binyamin Dissen <[email protected]> http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies. ---------------------------------------------------------------------- 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
