1 INFO-VAX	Tue, 27 Jun 2006	Volume 2006 : Issue 355       Contents:E Oops! (Was: Re: Stacks, Static, Item Lists, Modes and Initialization) I Re: Oops! (Was: Re: Stacks, Static, Item Lists, Modes and Initialization) # OT: Intel quad core X64 benchmarked  Proxy Server for OpenVMS?  Re: Proxy Server for OpenVMS?  Re: Proxy Server for OpenVMS? % Re: Security Setup and Product Advice  Re: Vax Debugger Examples  Re: Vax Debugger Examples  Re: Vax Debugger Examples   F ----------------------------------------------------------------------  % Date: Tue, 27 Jun 2006 20:37:43 +0800 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> N Subject: Oops! (Was: Re: Stacks, Static, Item Lists, Modes and Initialization)1 Message-ID: <e7r8n2$n72$1@news-02.connect.com.au>    Hi John,  H Thanks for the details of what BLISS actually does. If you can help withI what MACRO was doing with the following scenario (apart from Pilot-Error, H and no compiler can save complete idiots from themselves :-) then pleaseG jump in. You see, I made a boo-boo in the little example I was using to K check user-mode access to the exec-mode stack. I thought I was grabbing ten J pageLETS of stack space for some local storage and frigging around, when IK was actually taking ten pages. I'll have to find one of those lovely memory K layout diagrams with the "You are here" arrows 'cos at -81920(ExecSP) was a J usable page of memory with UREW protection that was happy to masquerade as the exec-mode stack.  I Anyway, it doesn't matter, and the main point is that the Exec-Mode stack K looks to default to about twenty-odd pagelets and is, in fact, protected at  Exec-RW and User-Nada access.   J Static compiler allocated Psects and the resulting PROTECTed sections thatJ the linker allocates are, on the other hand, still UREW. I quite like thatJ functionality and make use of it by passing Read-Only parameters to User'sH code. But this Local-Storage, Self-Cleansing Protected Stack bizzo looksE like the mut's nuts for UWSS house-keeping. Simply RETurn and it just  knows - spooky!    Cheers Richard Maher  K PS. Thanks again Glen for the "I think you're talking bollocks!" e-mail :-)   J PPS. Has anyone got some sort of example/tutorial (documentation pointer?)L on $setstk? Can (in combo) $setprt o/ride all of what's just been said? (NotH that I ever intend to do that, but as a backgrounder it could be useful)  3 "John Reagan" <john.reagan@hp.com> wrote in message , news:lgSng.2239$ZG2.1548@news.cpqcorp.net... > Richard Maher wrote: > I > > Now I'm not too bright with Bliss but my mate assures me that "Local"  goesE > > on the Stack and "Own" looks like static compiler generated data.  > H > I don't know anything about the code of which you speak, but I do know > BLISS. > L > OWN is indeed static, local storage.  LOCAL is indeed stack-based storage. > J > The INITIAL and PRESET attributes tell BLISS to initialize the variablesE > when they are created.  For OWN, that is done conceptually at image J > activation time although the initialization was probably done at compileF > or link time.  For LOCAL, each time the routine containing the LOCAL; > (with INITIAL or PRESET), that variable gets initialized.  > E > The contents of local variables when a routine exits is outside the J > definition of BLISS (or C for that matter).  If your BLISS or C tries toG > zero out a local variable as the last operation in a routine, you are C > likely to find the compiler's optimizer removing the operation as  > unnecessary. >  > --  
 > John Reagan 1 > HP Pascal/{A|I}MACRO for OpenVMS Project Leader  > Hewlett-Packard Company    ------------------------------  # Date: Tue, 27 Jun 2006 14:55:51 GMT & From: John Reagan <john.reagan@hp.com>R Subject: Re: Oops! (Was: Re: Stacks, Static, Item Lists, Modes and Initialization)2 Message-ID: <XFbog.2273$hi3.1132@news.cpqcorp.net>   Richard Maher wrote:
 > Hi John, > J > Thanks for the details of what BLISS actually does. If you can help withK > what MACRO was doing with the following scenario (apart from Pilot-Error, J > and no compiler can save complete idiots from themselves :-) then pleaseI > jump in. You see, I made a boo-boo in the little example I was using to M > check user-mode access to the exec-mode stack. I thought I was grabbing ten L > pageLETS of stack space for some local storage and frigging around, when IM > was actually taking ten pages. I'll have to find one of those lovely memory   G As you have found out, only the user-mode stack automatically expands.  F The other stacks have fixed sizes (see KSTACKPAGES and EXECSTACKPAGES  SYSGEN parameters).   F Even allocating 10 pagelets with 'SUBL2 #10*512, SP' is quite a bit ofF execstack space when the default EXECSTACKPAGES on my Alpha system is  only 2 8K pages.   --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  # Date: Tue, 27 Jun 2006 17:41:11 GMT ( From: Alan Greig <greigaln@netscape.net>, Subject: OT: Intel quad core X64 benchmarked> Message-ID: <X4eog.488643$tc.487563@fe2.news.blueyonder.co.uk>  D With the dual core Montecito finally about to appear, it seems that E quad-core X64 Intel samples are out and about. There's a benchmark at D http://www.xtremesystems.org/forums/showthread.php?p=1527913 if you  haven't yet seen it. --  
 Alan Greig   ------------------------------  + Date: Tue, 27 Jun 2006 07:56:43 +0000 (UTC) < From: gartmann@nonsense.immunbio.mpg.de (Christoph Gartmann)" Subject: Proxy Server for OpenVMS?) Message-ID: <e7qobr$f2f$1@news.BelWue.DE>    Hello,  J we have some PCs that have non-routed IP-addresses like 192.168.1.5 or so.H These PCs shouldn't be able to connect to the Internet. But they do haveF an anti virus software that needs updates. A similar situation is withK "Windows-Update" or "Office-Update". Our idea is now to do this via a proxy I server. Thus, is there something available to achieve this under OpenVMS? & We have Multinet V5.1 if that matters.   Regards,    Christoph Gartmann    --  E  Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452   ImmunbiologieI  Postfach 1169                 Internet: gartmann@immunbio dot mpg dot de   D-79011  Freiburg, Germany 9                http://www.immunbio.mpg.de/home/menue.html    ------------------------------  # Date: Tue, 27 Jun 2006 09:14:01 GMT L From: winston@SSRL.SLAC.STANFORD.EDU (Alan Winston - SSRL Central Computing)& Subject: Re: Proxy Server for OpenVMS?6 Message-ID: <00A57D17.D218D07C@SSRL.SLAC.STANFORD.EDU>  h In article <e7qobr$f2f$1@news.BelWue.DE>, gartmann@nonsense.immunbio.mpg.de (Christoph Gartmann) writes: >Hello,  > K >we have some PCs that have non-routed IP-addresses like 192.168.1.5 or so. I >These PCs shouldn't be able to connect to the Internet. But they do have G >an anti virus software that needs updates. A similar situation is with L >"Windows-Update" or "Office-Update". Our idea is now to do this via a proxyJ >server. Thus, is there something available to achieve this under OpenVMS?' >We have Multinet V5.1 if that matters.   K VMS-based webservers offer some proxy service.  I'm currently using a proxy K server in Apache (locked down to specific IP addresses) to allow updates of D a firewall appliance not otherwise visible on the internet.  I thinkM Windows-Update and Office-Update go over port 80 (don't know for sure); maybe  this would work.   -- Alan    ------------------------------  % Date: Tue, 27 Jun 2006 13:28:17 +0400 N From: "Ruslan R. Laishev" <zzLaishev@zzDeltaTelecom.RU-remove.all-zz-to-reply>& Subject: Re: Proxy Server for OpenVMS?? Message-ID: <504226FB8C043B8A11F159AD0B01858A@NNTP.DeltaTel.RU>    Hello, Christoph!    	We have used The WASD.    Christoph Gartmann wrote:  > Hello, > L > we have some PCs that have non-routed IP-addresses like 192.168.1.5 or so.J > These PCs shouldn't be able to connect to the Internet. But they do haveH > an anti virus software that needs updates. A similar situation is withM > "Windows-Update" or "Office-Update". Our idea is now to do this via a proxy K > server. Thus, is there something available to achieve this under OpenVMS? ( > We have Multinet V5.1 if that matters. > 
 > Regards, >    Christoph Gartmann  >    --  F + WBR, OpenVMS [Sys|Net] HardWorker ............. Skype: SysMan-One  +9 Delta Telecom JSC, IMT-MC-450(CDMA2000) cellular operator E Russia,191119,St.Petersburg,Transportny per. 3 Cel: +7 (812) 716-3222 F +http://starlet.deltatelecom.ru ............. Frying on OpenVMS only +   ------------------------------    Date: 27 Jun 2006 06:46:50 -0700% From: "ou8125150" <subasic@gmail.com> . Subject: Re: Security Setup and Product AdviceC Message-ID: <1151416010.265777.179750@b68g2000cwa.googlegroups.com>   G I just want to thank the folks who have responded to me with rock-solid G information and great advice.  You've given me some excellent resources A and a great start at finding the information that I need.  Thanks  again!   Bob Gezelter wrote:  > ou8125150 wrote:K > > I am a VMS newcomer and am seeking advice on where to look for security E > > lockdown advice, whitepapers, and/or products.  I am specifically  > > looking to be able to: > > J > >   1. Have very granular lockdown capabilities such that if someone hasG > > access to the $ prompt I can allow or deny access to run particular  > > images or commands> > >   2. Log and playback entire login (telnet based) sessionsH > >   3. Automatically disable accounts that have n (say 3 or 4) failed, > > consecutive login attempts. K > >   4. Have the system generate and send reports based on security events  > > in an admin-definable way. > > J > > >From what little I have gleaned from my reading, some or most of thisJ > > can be accomplished via standard VMS tools.  It appears that I may notE > > get all I need without a separate (perhaps commercial) tool.  Any ' > > advice will be greatly appreciated.  > F > OpenVMS has the inherant ability to control access in extremely fine? > granularity. If need be, this can be on a file by file basis. = > Thankfully, this degree of granularity is not often needed.  > F > My presentation "Building Secure Applications on OpenVMS" (slides atB > http://www.rlgsc.com/cets/2000/444.html ) gives some examples ofI > solutions and issues. Another presentation, "OpenVMS User Environments" ! > (a summary of the session is at B > http://www.rlgsc.com/hpworld/2004/N227.html ). That presentationH > focused on how privileges can be delegated, without requiring users to; > be granted privileges which could compromise other users.  > G > As Hoff and Larry have mentioned, the best general place to start for G > the basic information is the standard Guide to System Security, which G > is available in the OpenVMS documentation kit (on hardcopy, CDROM, or G > on the www via HP at http://www.hp.com/go/openvms ). Beyond that, the C > most productive method starts with understanding the security and H > integrity requirements of the environment, and determining the details% > of implementing these requirements.  > I > I hope that the above is helpful. If I have been unclear, please let me  > know.  > & > - Bob Gezelter, http://www.rlgsc.com@ >   Author, "OpenVMS Security", Handbook of Information Security$ >   Contributing Editor, OpenVMS.org   ------------------------------    Date: 27 Jun 2006 05:50:27 -0700$ From: "Chris L" <clusardi2k@aol.com>" Subject: Re: Vax Debugger ExamplesC Message-ID: <1151412627.120907.224610@u72g2000cwu.googlegroups.com>   / How would this change if I use an Ada compiler?   
 Thank you, Christopher Lusardi    briggs@encompasserve.org wrote: k > In article <1151334024.931867.30400@p79g2000cwp.googlegroups.com>, "Chris L" <clusardi2k@aol.com> writes: G > > Is there a link that can show me how to debug programs? A link that H > > shows how to step through a program, how to set break points, how toH > > set program variables, etc. I'm looking for all the goodies that the > > vax debugger does. > 
 > Example: >  > $ type test.c 	 > main ()  > {  >    int a;  >    a = 1;  >    a = a + a;  > }  > $ cc test /deb /noopt  > $ link test /deb
 > $ r test > 1 >          OpenVMS Alpha Debug64 Version V7.2-01R  > - > %DEBUG-I-INITIAL, Language: C, Module: TEST 3 > %DEBUG-I-NOTATMAIN, Type GO to reach MAIN program  > 	 > DBG> go  > break at routine TEST\main >      4:    a = 1;  > DBG> type 1:6 
 > module TEST  >      1: main ()  >      2: {  >      3:    int a;  >      4:    a = 1;  >      5:    a = a + a;  >      6: }  > DBG> set break %line 5 > DBG> g > break at TEST\main\%LINE 5 >      5:    a = a + a; 
 > DBG> e a > TEST\main\a:    1  > DBG> step  > stepped to TEST\main\%LINE 6 >      6: } 
 > DBG> e a > TEST\main\a:    2  > DBG> gJ > %DEBUG-I-EXITSTATUS, is '%SYSTEM-S-NORMAL, normal successful completion' > DBG> help  >  > DEBUG  > B >      The Debugger Command Dictionary contains detailed referenceE >      information about all debugger commands, organized as follows:  > A >      o  Command Format explains how to enter debugger commands.  > A >      o  Commands Disabled in DECwindows lists commands that are @ >         disabled in the command/message view of the debugger's% >         DECwindows Motif interface.  > F >      o  Messages gives general information about debugger diagnosticE >         messages. The Debugger Command Dictionary contains detailed < >         reference information about the debugger commands. >  >  >  > % >   Additional information available:  > @ >   New_Features          Address_Expressions   Built_in_SymbolsJ >   Client-Server_Interface          Command_Format        Command_Summary9 >   Debugging_Configurations         DECwindows_Interface : >   Commands_Disabled_in_DECwindows  Keypad_Definitions_CIH >   Keypad_Definitions_GUI           Language_Support      Logical_NamesK >   Messages   Path_Names SS$_DEBUG  System_Management     VWS_Workstations M >   @_(Execute_Procedure) ACTIVATE   ATTACH     CALL       CANCEL     CONNECT M >   Ctrl       DEACTIVATE DECLARE    DEFINE     DELETE     DEPOSIT    DISABLE N >   DISCONNECT DISPLAY    DO         DUMP       EDIT       ENABLE     EVALUATEH >   EXAMINE    EXIT       EXITLOOP   EXPAND     EXTRACT    FOR        GOL >   HELP       IF         MONITOR    MOVE       PTHREAD    QUIT       REBOOTL >   REPEAT     RERUN      RUN        SAVE       SCROLL     SEARCH     SELECTF >   SET        SHOW       SPAWN      STEP       SYMBOLIZE  SYNCHRONIZE >   TYPE       WHILE >  > DBG> exit  > $  >  >  > 
 > End example  > F > One command that I like but which doesn't come across at all well in > a text posting is: >  > DBG> SET MODE SCREEN > F > You get a three window display with the source code up top, debugger8 > output in the middle and debugger input in the bottom. > E > In screen mode, program output is presented in the same window with  > debugger input.  > H > In my experience, EXAMINE, STEP, SET BREAK and GO will do about 95% of/ > what you need done.  You can get fancier with  >  > STEP /INTO > STEP /RETURN > SET WATCH  > SET BREAK /AFTER) > SET BREAK <location> WHEN ( condition ) $ > SET BREAK <location> DO ( action ) > J > And you can also interrupt execution with control-Y and put the debuggerI > into control with $ DEBUG.  This is handy if you put your breakpoint at H > the wrong place, if you want to see how far your program has gotten or' > if you're debugging an infinite loop.  > 
 > Example: >  > $ type test.for  >         integer i  >         i = 0  > 10      i = i + 1  >         goto 10 
 >         end  > $ fort test /deb /noopt  > $ link test /deb > $ r 
 > _File: test  > 1 >          OpenVMS Alpha Debug64 Version V7.2-01R  > 8 > %DEBUG-I-INITIAL, Language: FORTRAN, Module: TEST$MAIN >  > DBG> g >  Interrupt > 	 > $ debug 
 > DBG> e i > TEST$MAIN\I:    115978496  > DBG>  Exit   ------------------------------    Date: 27 Jun 2006 09:52:33 -0500 From: briggs@encompasserve.org" Subject: Re: Vax Debugger Examples3 Message-ID: <ZVwSH$6U2cdw@eisner.encompasserve.org>   j In article <1151412627.120907.224610@u72g2000cwu.googlegroups.com>, "Chris L" <clusardi2k@aol.com> writes:1 > How would this change if I use an Ada compiler?   E It's been quite a while since I've compiled an Ada program, but, as I   recall, it's along the lines of:   $ TYPE TEST.ADA 
 with text_io;  procedure test is  begin #    text_io.put ( "Hello, world!" );  end;- $ ACS CREATE LIB [.ADALIB]	! One time command 9 $ ACS SET LIB [.ADALIB]		! Need this each time you log in  $ ADA TEST.ADA /DEBUG /NOOPT $ ACS LINK TEST /DEBUG
 $ RUN TEST  & and then the debugger pops up as usual  @ I think you can do the compilation step from within ACS (the Ada Compiler System) if you wish.   : Larry Kilgallen is sure to know the details better than I.   ------------------------------    Date: 27 Jun 2006 10:03:20 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) " Subject: Re: Vax Debugger Examples3 Message-ID: <KpavhNOYop+Z@eisner.encompasserve.org>   j In article <1151412627.120907.224610@u72g2000cwu.googlegroups.com>, "Chris L" <clusardi2k@aol.com> writes:  1 > How would this change if I use an Ada compiler?   D If you are using VAX/DEC/Compaq/HP Ada, you should look for a manualE called "Developing Ada Programs on VMS Systems".  The order number on E the hardcopy one I have is AA-EF86B-TE, but that is from May of 1989, B I am sure there are more recent versions, but at the level you are/ asking about, even one that old should be good.    ------------------------------   End of INFO-VAX 2006.355 ************************