1 INFO-VAX	Mon, 07 Aug 2006	Volume 2006 : Issue 437       Contents:= Re: "OpenVMS - A System of Structure" published on OSNEWS.COM 5 Re: Authentication for outgoing TCP/IP Services mail? 5 Re: Authentication for outgoing TCP/IP Services mail? 5 Re: Authentication for outgoing TCP/IP Services mail? 5 Re: Authentication for outgoing TCP/IP Services mail? 5 Re: Authentication for outgoing TCP/IP Services mail? 5 Re: Authentication for outgoing TCP/IP Services mail? 5 Re: Authentication for outgoing TCP/IP Services mail? 5 Re: Authentication for outgoing TCP/IP Services mail? 5 Re: Authentication for outgoing TCP/IP Services mail? & Re: JAVA: The minimum you need to know& Re: JAVA: The minimum you need to know OpenVMS V8.3 Doc's Online  Re: Speaking of Clusters: ) Re: strange problem: satellite won't boot  Re: USB numeric keypad anyone? Re: USB numeric keypad anyone? Re: USB numeric keypad anyone?6 Re: VMS>7.3 Linker Manual - I64 Short Data restriction6 Re: VMS>7.3 Linker Manual - I64 Short Data restriction  F ----------------------------------------------------------------------  $ Date: Mon, 7 Aug 2006 06:47:11 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> F Subject: Re: "OpenVMS - A System of Structure" published on OSNEWS.COM< Message-ID: <44d71926$0$24211$9a6e19ea@news.newshosting.com>  5 "Bob Gezelter" <gezelter@rlgsc.com> wrote in message  < news:1153234320.846189.73730@m79g2000cwm.googlegroups.com...@ > OSNews.com has just published an article I wrote on OpenVMS at/ > http://www.osnews.com/story.php?news_id=15222  > & > - Bob Gezelter, http://www.rlgsc.com >   < For some reason I only discovered this article today (Aug-7)- http://www.osnews.com/story.php?news_id=15222 5 but it is kind of cool to see OpenVMS being promoted.   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html9 http://www3.sympatico.ca/n.rieck/links/openvms_demos.html    ------------------------------   Date: 07 Aug 2006 10:38:37 GMT0 From: spud_demon@e.THUNDERMAKER.NET (Spud Demon)> Subject: Re: Authentication for outgoing TCP/IP Services mail?6 Message-ID: <44d7182d$0$42180$815e3792@news.qwest.net>   "David J. Dachtera" <djesys.no@spam.comcast.net> writes in article <44D645C4.E45079FC@spam.comcast.net> dated Sun, 06 Aug 2006 14:40:52 -0500:7 >My take on his question was similar to my own quandry:  > E >From my hobbyist Alpha, how to I MAIL something to an SMTP recipient = >using my ISP's SMTP agent as the up-line forwarder when SMTP " >authentication is being enforced?  F Why tie yourself to your ISP's mail server?  OpenVMS has one, and it's? fully capable of routing your messages to the right MX domains.   L Keeping messages on your own system has the advantage that you can watch the8 queue and know if your buddy's server is not responding.  = >A mechanism similar to this would be helpful, if it existed:  > 8 >$ DEFINE/SYSTEM/EXEC SMTP$SEND_AUTH "djesys:mypassword"  L Somebody would leave an open relay, get their account revoked, and blame HP.  . --Spud Demon		spud_demon -at- thundermaker.net   ------------------------------  * Date: Mon, 7 Aug 2006 11:27:04 +0000 (UTC) From: david20@alpha2.mdx.ac.uk> Subject: Re: Authentication for outgoing TCP/IP Services mail?, Message-ID: <eb7828$c94$1@south.jnrs.ja.net>  i In article <44d7182d$0$42180$815e3792@news.qwest.net>, spud_demon@e.THUNDERMAKER.NET (Spud Demon) writes:  >"David J. Dachtera" <djesys.no@spam.comcast.net> writes in article <44D645C4.E45079FC@spam.comcast.net> dated Sun, 06 Aug 2006 14:40:52 -0500: 8 >>My take on his question was similar to my own quandry: >>F >>From my hobbyist Alpha, how to I MAIL something to an SMTP recipient> >>using my ISP's SMTP agent as the up-line forwarder when SMTP# >>authentication is being enforced?  > G >Why tie yourself to your ISP's mail server?  OpenVMS has one, and it's @ >fully capable of routing your messages to the right MX domains. > M >Keeping messages on your own system has the advantage that you can watch the 9 >queue and know if your buddy's server is not responding.  > I Sending mail directly when you have a dynamic address is not a good idea. L Too many mail servers use blacklists such as MAPS DUL or SORBS DUHL to blockJ mail from such addresses. Hence it is better to send mail out through your2 ISP's central mailhubs which shouldn't be blocked.N This can usually be accomplished by either configuring your client to send outI directly through your ISP's mailhub or by configuring your mail server to F forward all of it's mail out through the ISP's mailhub as a smarthost.  K However if your ISP's mailhub forces authentication then the latter becomes J impossible unless your mailserver software has some means of providing the authentication when challenged.   ' The only solutions in that case are to    I 1) Get such functionality added to the mailserver software which is what      David wants   or  C 2) Use a different system as a smarthost which doesn't require such N    authentication. A number of providers such as dyndns provide such a service    at a price.      > >>A mechanism similar to this would be helpful, if it existed: >>9 >>$ DEFINE/SYSTEM/EXEC SMTP$SEND_AUTH "djesys:mypassword"  > M >Somebody would leave an open relay, get their account revoked, and blame HP.  >   O If you are running your own mailserver then you are responsible for making sure N it isn't an open-relay. If you are running an open-relay and your ISP suspends7 or closes your account then that is your fault not HPs.       
 David Webb Security team leader CCSS Middlesex University            / >--Spud Demon		spud_demon -at- thundermaker.net    ------------------------------  % Date: Mon, 07 Aug 2006 15:08:58 +0200 / From: Paul Sture <paul.sture.nospam@hispeed.ch> > Subject: Re: Authentication for outgoing TCP/IP Services mail?; Message-ID: <4075c$44d73b6b$50db5015$17930@news.hispeed.ch>    david20@alpha2.mdx.ac.uk wrote: k > In article <44d7182d$0$42180$815e3792@news.qwest.net>, spud_demon@e.THUNDERMAKER.NET (Spud Demon) writes:  >> "David J. Dachtera" <djesys.no@spam.comcast.net> writes in article <44D645C4.E45079FC@spam.comcast.net> dated Sun, 06 Aug 2006 14:40:52 -0500: : >>> My take on his question was similar to my own quandry: >>>    Correct David.  H >> >From my hobbyist Alpha, how to I MAIL something to an SMTP recipient@ >>> using my ISP's SMTP agent as the up-line forwarder when SMTP% >>> authentication is being enforced? I >> Why tie yourself to your ISP's mail server?  OpenVMS has one, and it's B >> fully capable of routing your messages to the right MX domains. >>O >> Keeping messages on your own system has the advantage that you can watch the ; >> queue and know if your buddy's server is not responding.  >>K > Sending mail directly when you have a dynamic address is not a good idea. N > Too many mail servers use blacklists such as MAPS DUL or SORBS DUHL to blockL > mail from such addresses. Hence it is better to send mail out through your4 > ISP's central mailhubs which shouldn't be blocked.  H I have reason to believe that my ISP has in the past suffered from spam G zombies unwittingly run by their customers, so IMHO they did the right  " thing by enforcing authentication.  G In any case I have found at least one place where dynamic addresses in  ( my range are, or have been, blacklisted.  G Since I am now using cable, and don't need a telco line any more, JF's  D suggestion of getting a different ISP would automatically double my D monthly charges for little gain (and at that cost, I'd still have a > dynamic address, and could run into exactly the same problem).  4 For me, it's just a "would be nice to have" feature.  I One alternative I can think of is to use my OS X system to act as a mail  B server which can forward from my VMS system using the appropriate 4 authentication, but I've done zero research on that.   --    
 Paul Sture   ------------------------------  # Date: Mon, 07 Aug 2006 13:53:58 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>> Subject: Re: Authentication for outgoing TCP/IP Services mail?2 Message-ID: <WBHBg.1633$qs4.1518@news.cpqcorp.net>   Paul Sture wrote:   6 > For me, it's just a "would be nice to have" feature.  G    I've passed along a question on this (basically ESMTP SMTP-MTA port  K 587 submission via the SMTP transport image), and a pointer to this thread.   E    SMTP servers (MTAs; which is what TCP/IP Services provides) don't  + send their submissions traffic to port 587.   F    The other approaches are to seek and acquire ISP authorization for C your SMTP mail server, or to use an alternate mail client (MSA) to   submit your email.   ------------------------------  * Date: Mon, 7 Aug 2006 14:09:22 +0000 (UTC) From: david20@alpha2.mdx.ac.uk> Subject: Re: Authentication for outgoing TCP/IP Services mail?, Message-ID: <eb7hii$f1m$1@south.jnrs.ja.net>  a In article <WBHBg.1633$qs4.1518@news.cpqcorp.net>, Hoff Hoffman <hoff-remove-this@hp.com> writes:  >Paul Sture wrote: > 7 >> For me, it's just a "would be nice to have" feature.  > H >   I've passed along a question on this (basically ESMTP SMTP-MTA port L >587 submission via the SMTP transport image), and a pointer to this thread. > F >   SMTP servers (MTAs; which is what TCP/IP Services provides) don't , >send their submissions traffic to port 587. > G >   The other approaches are to seek and acquire ISP authorization for  D >your SMTP mail server, or to use an alternate mail client (MSA) to  >submit your email.   H I'd raised this issue on the vmsnet.mail.pmdf newsgroup when PMDF becameD available to hobbyists and have now formally raised it with Process ) (through our local reseller - Essential).     
 David Webb Security team leader CCSS Middlesex University   ------------------------------   Date: 7 Aug 2006 09:50:59 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) > Subject: Re: Authentication for outgoing TCP/IP Services mail?3 Message-ID: <CkokGTSaXQYE@eisner.encompasserve.org>   m In article <415a8$44d4c9ff$50db5015$27406@news.hispeed.ch>, Paul Sture <paul.sture.nospam@hispeed.ch> writes: K > Some time ago, my ISP changed their mail system so that I need to supply  H > authentication for outgoing mail (Good Move against spamming zombies). > A > In Mac Mail, Thunderbird etc, there are settings for supplying  # > authentication for outgoing mail.  > 0 > Does similar exist in TCP/IP Services for VMS? >   F    That kind of thing can be fonud in most browsers.  Try reading your#    email via Mozilla, works for me.    ------------------------------  * Date: Mon, 7 Aug 2006 15:04:14 +0000 (UTC) From: david20@alpha2.mdx.ac.uk> Subject: Re: Authentication for outgoing TCP/IP Services mail?, Message-ID: <eb7kpe$fuv$1@south.jnrs.ja.net>  q In article <CkokGTSaXQYE@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: n >In article <415a8$44d4c9ff$50db5015$27406@news.hispeed.ch>, Paul Sture <paul.sture.nospam@hispeed.ch> writes:L >> Some time ago, my ISP changed their mail system so that I need to supply I >> authentication for outgoing mail (Good Move against spamming zombies).  >>  B >> In Mac Mail, Thunderbird etc, there are settings for supplying $ >> authentication for outgoing mail. >>  1 >> Does similar exist in TCP/IP Services for VMS?  >>   > G >   That kind of thing can be fonud in most browsers.  Try reading your $ >   email via Mozilla, works for me. >   E True enough but doesn't help if you want to run your own mail server.       
 David Webb security team leader CCSS Middlesex University   ------------------------------  # Date: Mon, 07 Aug 2006 16:31:57 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>> Subject: Re: Authentication for outgoing TCP/IP Services mail?1 Message-ID: <1WJBg.1645$tG4.883@news.cpqcorp.net>    david20@alpha2.mdx.ac.uk wrote:   G > True enough but doesn't help if you want to run your own mail server.   (    SMTP mail servers (MTAs) use port 25.  H    SMTP mail clients (MUAs) use port 25 or (with SMTP-MTA ESMTP RFC2476 * support) to the MSA listening on port 587.  F    I've suggested a modification to the SMTP transport to allow it to H connect into port 587, as the SMTP transport image for the OpenVMS MAIL   utility (SMTP%) is the MUA here.  H    Basically either the SMTP transport client has to be modified to use G port 587, or there needs to be some sort of a translating gateway here  F that accepts connections on 25 and passes them along on 587.  Or some F other MAIL client need be used, or a custom SMTP MAIL transport image I need be created, or the ISP needs to designate this OpenVMS system as an   MTA.  (Which it is.)   ------------------------------  % Date: Mon, 07 Aug 2006 13:32:21 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> > Subject: Re: Authentication for outgoing TCP/IP Services mail?, Message-ID: <44D77924.FA40ECD9@teksavvy.com>   Hoff Hoffman wrote: G >    The other approaches are to seek and acquire ISP authorization for D > your SMTP mail server, or to use an alternate mail client (MSA) to > submit your email.  G You do realise that you are effgectively suggesting one move off VMS ?  ! ("use an alternate mail client").   D The point is that we want VMS to be modern and usable in the currentH environment. SMTP is being tightened up in many ways to prevent spam. SoF either HP re-hires engineers to do work on the TCPIP Services product,F or you open-source it NOW so that we (the users) can gradually add the missing features.    ------------------------------  $ Date: Mon, 7 Aug 2006 07:01:15 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> / Subject: Re: JAVA: The minimum you need to know < Message-ID: <44d71c72$0$24172$9a6e19ea@news.newshosting.com>  # <sean@obanion.us> wrote in message  = news:1154877430.972546.194510@h48g2000cwc.googlegroups.com... A > The server in the order link does not seem to be responding.... B > And the Java book is not listed on Island's reference book page. > 2 > Anybody have any success this weekend with this? >  >  > Sean >   H I can ping the server but port 80 appears to be dead (they probably use  Windows to take orders :-)  K Until they reboot their ordering software, I've posted a few comments here: U http://www3.sympatico.ca/n.rieck/links/cool_openvms.html#TMYNTK-About-Java-on-OpenVMS   K I've only read the "Introduction" and "Chapter 1" so far. The Introduction  K contains a humorous (but probably true) account about how IT decisions are  L made "On Mahogany Row". This partially explains why OpenVMS software people 1 are sometimes forced to implement Java solutions.     
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html: http://www3.sympatico.ca/n.rieck/links/openvms_demos.html    ------------------------------  $ Date: Mon, 7 Aug 2006 11:20:15 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> / Subject: Re: JAVA: The minimum you need to know < Message-ID: <44d75927$0$24198$9a6e19ea@news.newshosting.com>  # <sean@obanion.us> wrote in message  = news:1154877430.972546.194510@h48g2000cwc.googlegroups.com... A > The server in the order link does not seem to be responding.... B > And the Java book is not listed on Island's reference book page. > J Because there are only a limited (< 100) copies of the JAVA book, I don't I think this title is ever going to be listed on Island's OpenVMS Book Page  >  [...snip...] >  These links are now up: : https://www.icusc.com/NOV21_WEBSTORE.asp?search_fd2=*JAVA*E https://www.icusc.com (just search for JAVA in the description field)   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.! http://www3.sympatico.ca/n.rieck/    ------------------------------  $ Date: Mon, 7 Aug 2006 12:06:08 -0400' From: "Main, Kerry" <Kerry.Main@hp.com> " Subject: OpenVMS V8.3 Doc's OnlineT Message-ID: <FA60F2C4B72A584DBFC6091F6A2B86840187376F@tayexc19.americas.cpqcorp.net>   All,  D Readers of this list may be interested in the new OpenVMS V8.3 doc's that are now online at: - http://h71000.www7.hp.com/doc/os83_index.html   
 New Features: ? http://h71000.www7.hp.com/doc/83FINAL/6679/OVMS_83_FEATURES.PDF      Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------   Date: 7 Aug 2006 09:16:41 -0700 - From: "Andrew" <andrew_harrison@symantec.com> " Subject: Re: Speaking of Clusters:C Message-ID: <1154967401.657250.127190@h48g2000cwc.googlegroups.com>    Keith Parris wrote:  > Doug Phillips wrote:  > > curiosity leads me to ask ifI > > anyone here has experience with both Veritas and OpenVMS clusters and  > > how they compare?  > F > As I understand it, the basic capability that Veritas Cluster ServerI > provides is a cluster-wide file system, which is a prerequisite for any J > sort of failover capabilities within a cluster (e.g. you have to be ableJ > to see the files before you can start up an application which uses thoseM > files), and some support for application failover operations between nodes.  >   . I work for VERITAS now Symantec so I can help.  B Symantec provides a range of HA products starting from Filesystem,A Volume management and moving up through N+1 cluster failover to , F Cluster Volume Manager, Cluster Filesystem and support for Oracle RAC.  B You do not need the Cluster Filesystem to implement a cluster withF VERITAS Cluster Server, you can simply allow VCS to import the volumesG from the failed cluster node into the running node when it detects that D the production node has failed and then restart the apps. This worksB with VERITAS Volume manager and a range of other Volume Management	 products.   B Cluster Filesystem is an option which gives you multi-node reader,1 writer clustering, VCS is of course aware of CFS.    Ditto for Oracle RAC.   H > MC/Serviceguard did not have a cluster-wide file system before, so for3 > them it undoubtedly seemed the logical next step.  > G > What it does not provide is a Distributed Lock Manager, or any of the I > file-sharing and coordination functions within a cluster that are built  > on top of a DLM. > I > So access to disks is still on a one-node-at-a-time basis -- you'd need J > to scale up within an SMP box for more horsepower, since you can't scaleJ > out across multiple nodes running the same workload and sharing the sameK > files (except presumably for Oracle RAC, which has its own DLM built in).   F Err no, if you are using CFS then this is not the case, multiple nodes? can update the same filesystem at the same time. Consistency is 2 maintained usin VERITAS GLM (Global Lock Manager).  F CFS currently supports 32 same platform nodes, Solaris, AIX, HP-UX, orF Linux. It is not available for Windows although VCS and Volume Manager are.   Regards  Andrew Harrison    ------------------------------   Date: 7 Aug 2006 09:47:53 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) 2 Subject: Re: strange problem: satellite won't boot3 Message-ID: <kU6IxVj254Ux@eisner.encompasserve.org>   w In article <eb0ce1$tvg$1@online.de>, helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes:  > / > Any chance that PHONE will get IP capability?  >   H    The corresponding IP tool is usually known as talk.   Most people use(    IM for this now.  Why invent another?   ------------------------------  # Date: Mon, 07 Aug 2006 13:48:26 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>' Subject: Re: USB numeric keypad anyone? 1 Message-ID: <KwHBg.1632$tA4.256@news.cpqcorp.net>    Ken Fairfield wrote:  D >    So Hoff, do you have a terminal emulator that allows you to use? > all of the LK463 keys within EVE (under Windows XP or newer)?   B    I've been using the (available on the OpenVMS Freeware) VTstar I terminal emulator package for some eons now.  I haven't tried it with an  H LK463 however, and yet don't have one of those keyboards in my hardware I stash.  (I'll see if I can acquire one, and try it -- they're too new to  ( be showing up in my usual supply chain.)   ------------------------------  # Date: Mon, 07 Aug 2006 16:12:58 GMT " From:   VAXman-  @SendSpamHere.ORG' Subject: Re: USB numeric keypad anyone? 0 Message-ID: <00A59DA3.48F86242@SendSpamHere.ORG>  ` In article <KwHBg.1632$tA4.256@news.cpqcorp.net>, Hoff Hoffman <hoff-remove-this@hp.com> writes: >  >  >Ken Fairfield wrote:  > E >>    So Hoff, do you have a terminal emulator that allows you to use @ >> all of the LK463 keys within EVE (under Windows XP or newer)? > C >   I've been using the (available on the OpenVMS Freeware) VTstar  J >terminal emulator package for some eons now.  I haven't tried it with an   = Can you relay this to the person that provided this package:   How about a Mac OS X version?      --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  # Date: Mon, 07 Aug 2006 16:52:27 GMT , From: Hoff Hoffman <hoff-remove-this@hp.com>' Subject: Re: USB numeric keypad anyone? 1 Message-ID: <fdKBg.1647$PI4.977@news.cpqcorp.net>     VAXman- @SendSpamHere.ORG wrote:b > In article <KwHBg.1632$tA4.256@news.cpqcorp.net>, Hoff Hoffman <hoff-remove-this@hp.com> writes:  ? > Can you relay this to the person that provided this package:  ! > How about a Mac OS X version?     H    There's no current owner.  VTstar dates back to the Multia days, and A has been under ad-hoc maintenance since its retirement.  At best.   ?    An OS X port would require a hardware delivery, and several  , sequential deliveries of copious spare time.  H    The port of VTstar was most recently via the Microsoft Visual Tools; E it's not likely going to be a straight port over to OS X.  (One snag  C with that port and any with any currently-hypothetical source code  H release, too, is that the VTstar source code requires a custom-modified H and copyrighted Microsoft Visual C source code module.  That was one of : the central snags when last the release was contemplated.)  E    It might well be faster to start out with a DECterm port, as that  F already knows about X Windows, or to teach (for instance) an emulator F like PuTTY about an LK-series keyboard.  But that's all just theories 
 right now.  3    And yes, I know where the VTstar source code is.    ------------------------------  # Date: Mon, 07 Aug 2006 15:01:55 GMT & From: John Reagan <john.reagan@hp.com>? Subject: Re: VMS>7.3 Linker Manual - I64 Short Data restriction 1 Message-ID: <DBIBg.1639$sx1.741@news.cpqcorp.net>    Richard Maher wrote: > Hi,  > L > When could one expect to see a new version of the Linker Utility ReferenceM > Manual available with VMS. The one on the Web and Doc CDs appears to be VMS F > 7.3 April 2001. All the I64 stuff I've seen is in one of the 8.2 NewH > Features Manual (Which would make sense if 8.2 was the first supported> > itanium version :-) but are we gonna get a new set with 8.3?  - Yes, an updated linker manual comes with 8.3.    > K > Also, Can anyone shed any light of possible solutions/work-arounds to the I > 4MB limit on this Short Data segment thing? (Something other than "Your 1 > app's crap - re-engineer it!" would be nice :-)   D In general, filling up the 4MB short section results from compilers H asking for too many address constants.  Although the short section also D contains linker-created function descriptors and static user data 8  bytes or smaller.   I As background, on Alpha, each routine is passed a routine-unique pointer  I to the "linkage" section which contains its procedure descriptor and all  D the compiler requested address constants.  While there isn't a hard C limit, the downside is that if multiple compilations need the same  @ address, there are many copies in the final image.  On I64, the G compilers don't allocate space for address constants, but request that  G the linker materialize the address for them.  It makes the linker more  @ complicated, but it also allows the linker to remove duplicates.  D By convention (we didn't 'invent' the convention, it comes from the D Intel Run-Time Conventions), this linker-created section of address I constants (and function descriptors, etc.) is shared by the whole image.  E   Each routine has the address of this "short" section passed in R1.  > The short section is accessed with the following instructions:  B              ; Add offset from R1 to the location where the linkerI              ; allocated the address.  This form of the 'add' instruction                ; is defined to be:'              ;       add Ra = imm22, Rb 9              ; where Rb must be either R0, R1, R2, or R3.               ;F              ; When the compiler generates the instruction, it usuallyF              ; puts a 0 in the 22-bit literal and issues a relocation.F              ; The relocation tells the linker to create an address toD              ; SYM, allocate it in the short section, and modify theF              ; immediate operand of the instruction to have the offset0              ; from R1 to the allocated address.              ;%              add Rn = @ltoff(SYM), R1   D              ; Now that we have the address of the address, fetch it              ;              ld8 Rn = [Rn]  G              ; We can now fetch the static data.  It might be something F              ; readonly like a linker-materialized address or it might:              ; be some user variable (8 bytes or smaller).              ;5              ld8 Rn = [Rn]  ; fetch an 8-byte 'thing'    > N > Specifically what, if anything, does the Option File command "SHORT_DATA=NO"# > do for us? (Latent support only?)    Doesn't do anything for you.   > J > I have seen the /SEGMENT=(Short_data=WRITE) but another 65K is not gonnaF > help us and it would be could to avoid any behavioural changes where > possible.   C You can also try /SEGMENT=SYMBOL_VECTOR=NOSHORT (if you creating a  F shareable image), avoid /NONATIVE_ONLY (it makes function descriptors B larger), and avoid pulling code sections from the default cluster.7 However, I wouldn't expect much help from any of these.    > J > There seems to be a fair sprinkling of discussions on the web (Some withN > Linux and Windows) but it seems to come down to the compiler having to do it > or you're stuffed? > 0 > Will PSECT_ATTR=data,NOSHORT never be allowed? > N > Can we not CLUSTER/COLLECT out of the SHORT segment hand have the Linker fix9 > up the "little" address references to real 32-bit ones?   F As you can see from my description above, the compiler generated code G with explicit knowledge that the data in question (user static data or  G linker-created addresses/FDs) can be found from an offset from R1.  If  E you tell the linker not to put that PSECT "near" R1, the code cannot   possibly reach it.  F Given the model used, on I64 you can compile all your modules without > any guarantee that you can combine them into one applications.  C Right now, I'm working on a COBOL issue where it generates address  I constants for code labels to use with the PERFORM statement.  I have two  H customers who cannot link due to this.  I have prototype COBOL compiler D read for testing by one of the customers now.  I'm off for a 3 week G absense [2-wks vacation, 1-wk business trip].  I'll finish it when I'm   return.   C Now, for those of you who have read the Itanium ELF ABI to see the  D complete list of relocations defined, you might notice a handful of F relocations that would let the linker fill in the 64-bit literal of a F 'movl' (move with long literal) instruction.  Such a relocation would D eliminate the 22-bit (4MB) short section by just not having a short G section.  However, the compilers would have to generate different code  G and the linker would have to implement them.  I don't believe that any  E Itanium linker (Windows, Linux, HP-UX, OpenVMS) implement the larger  G relocations or do any compilers produce them.  It makes the code quite  E position dependent and would require a $h!t load of image activation  @ time relocations especially for shareable images built that way!  D Is your application in COBOL?  Other than COBOL, I've yet to see an D application that won't link on OpenVMS I64.  Of course, knowing the @ above information, it would be trivial to hand-construct (well, 5 machine-construct) an application that wouldn't link.    --   John Reagan 5 HP Pascal/{A|I}MACRO/COBOL for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------   Date: 7 Aug 2006 09:53:28 -0700   From: "Ian Miller" <ijm@uk2.net>? Subject: Re: VMS>7.3 Linker Manual - I64 Short Data restriction C Message-ID: <1154969608.619757.321800@b28g2000cwb.googlegroups.com>    The updated linker manual is at 7 http://h71000.www7.hp.com/doc/83final/4548/4548PRO.HTML    ------------------------------   End of INFO-VAX 2006.437 ************************                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               P-"@}2"ȱDdkYi/be'JYUibE+gE!}\לD*"O`ƽȀ4F LJB?(lm"ItO
BOxxlTtBN&t+] Z#Tҕ:ӔhLܷ̍wD)ߝ؍EtXH9Zސ0qM_:)QPnEҥC?e㠖)˜-"|&߲Er&fie62mp_RXcnѣkL$|r']8AGvqD`z-eJ5%#N	ȈUdD,KQ.ޛ5C9TB{	~NIPrPD"D&9ް2)JG4jfi|7"zb>M_jwe0WeYyq^\D_`oXOi"w
Yu_hzS%L`Pƴ4[/4*6=}:ߞ<Ȕb&WtmD5aE_ _%8A~OFG͌)2#w{\!w{)~_ҏt@TvVpLzP)WtJmYj7EU?BA6Y-@=Jy zbQlDuIjVR/T\u?j<SW|S1
>0/) ڋ둬 Ih n9lUk7JHt;M<lsb,-CEV~垗  jl2B*E4  v'60OGĬQD/HcWM|q8aTZh7lȮ1¸M΀5
L-+$ /mN&eV^XQt^F*_adnr?s
jRXⒼ4(]:J-zwZQDL[6v>
ٜ7ٕB/ՙe=G
e8,%Kq_ +dm,'x,+om>ZR1=εMlS_lt}Ygk
{e+Q̨5ɢ<J .eQ1L.Jq΅.߂kô}/{Zsupl=aV%
tix{VMhXQ? Q#^
\?/AgM_&UlqCՁ՟1TğI2n@G堠vPɎP:RKVYP//zסI=޺;) 20DH,^:kY@~|idл}?Rh6<[d䭯ެ5Dq2Cռ1bo&y)2(}f+]1Ɓ9VX cdM@Fр:!FǾEb*lOKg B/&ϊ'T{QkoWF:쌠W{r&j rA83tYo/ W<=䝆	Rut?f~W܋WfB8Ѷgf4`7aa
"YEwi(Q
CDvH\&d̪!v
V+H#W._;o>*}|
!H gJ
z<2=|:ZY=2gڼەX +(*rQ68FD穋β=q
7Џ+Hf5--^V3  qz~%@D?qUmQʆ xT2XZ|$KSLX'	,ͬCqޗ"@ʇef~
4?%?~Okӵ(zh7q8*" P48ŵ-ecdVt):#{=K|.(GgJn` Ԩ<sC!H.r^q
tNMwYUly#}#;/9& ?dAɠqZ%nƫ o	,o.a:ާ*kr*SsC։!LoK< ):M4`(*8(xDT0x
kZM^qύeHtIz׌-cn	u_䴲NTE2&4K HV1c[T	$C8'7itXsGݥjڐh-rxT/[̖@MٓLޙeD17ΰ)?BJ r,
dy
MszSN`xv6E}4V5UUE:vRgRea\ԡh^	OQ'@$y^)?bHpp<^
J`wʤ-OuyD.3r&otl0Y!s@-RwKH
!C#qUUa6qNk.te0 Dhd{ȸ}Z
1ˤ`3ںA7>6%#_!/oHi5ٜWh.tl%D\۠nM\C_2<W@xBƦ<ac[kqw`'mv۶mضvcnFMdzzȑ5{%^RKdUw]jT4̥xAviB5X.ehIA
Ï_*/U^@ J`7_G}	VTߤRpV6|}4[3x?z$ڀ?wL<]+|J/4