1 INFO-VAX	Sun, 19 Dec 2004	Volume 2004 : Issue 702       Contents:0 Re: AUDIT_SERVER stops sending to the /LISTENER?0 Re: AUDIT_SERVER stops sending to the /LISTENER?! Re: DS10 on a NETGEAR DHCP Router  Heirarchal network question  Re: Heirarchal network question  Re: HP exodus to Intel Re: HP exodus to Intel Re: HP exodus to Intel Re: HP exodus to Intel Re: Interesting coding tidbit  Re: soon to be gone  Re: Time to revive Emerald?  Re: Time to revive Emerald?  Re: Time to revive Emerald? , Re: Vanquishing the Smoke and Fog with Facts Re: VESA/VGA BIOS emulation  Re: VESA/VGA BIOS emulation  Re: VMS with BULL terminals   F ----------------------------------------------------------------------  % Date: Sat, 18 Dec 2004 20:55:30 +0100 * From: Paul Sture <nospam@sture.homeip.net>9 Subject: Re: AUDIT_SERVER stops sending to the /LISTENER? , Message-ID: <32jg9cF3nl5m0U1@individual.net>   John Gemignani, Jr. wrote:B > "David J Dachtera" <djesys.nospam@comcast.net> wrote in message ' > news:41C0E5D8.EB4FB074@comcast.net...  >  >>"John Gemignani, Jr." wrote: >>L >>>For some reason today the audit server stopped sending sysuaf updates to  >>>theL >>>listener process.  I also no longer see the opcom alarms as well.  I did  >>>not6 >>>make any changes to the system to make this happen. >>> ' >>>Here are the current audit settings:  >>>  >>>$ sho audit/all >>>List of audit journals:% >>>  Journal name:           SECURITY 3 >>>  Journal owner:          (system audit journal) G >>>  Destination:            DISK$LOGS:[SECURITY]SECURITY.AUDIT$JOURNAL $ >>>  Monitoring:             enabledD >>>    Warning thresholds,   Block count:    100     Duration:    2 
 >>>00:00:00.0 D >>>    Action thresholds,    Block count:     25     Duration:    0 
 >>>00:30:00.0  >>> , >>>Security auditing server characteristics:  >>>  Database version:       4.4* >>>  Backlog (total):        100, 200, 300! >>>  Backlog (process):      5, 2 ! >>>  Server processing intervals: * >>>    Archive flush:        0 00:01:00.00* >>>    Journal flush:        0 00:05:00.00* >>>    Resource scan:        0 00:05:00.006 >>>  Final resource action:  purge oldest audit events >>> " >>>Security archiving information:! >>>  Archiving events:       none  >>>  Archive destination:  >>> 0 >>>System security alarms currently enabled for: >>>  ACL >>>  Authorization >>>  Audit:         illformed 8 >>>  Breakin:       dialup,local,remote,network,detached >>>  Logfailure:? >>>batch,dialup,local,remote,network,subprocess,detached,server  >>> 0 >>>System security audits currently enabled for: >>>  ACL >>>  Authorization >>>  Audit:         illformed 8 >>>  Breakin:       dialup,local,remote,network,detachedI >>>  Logfailure:    batch,dialup,local,remote,network,subprocess,detached  >>> $ >>>I don't get it.  Why did it stop? >>G >>If you're not getting OPCOM messages, it's fair bet the OPCOM process H >>has a problem. Depending what state OPCOM is in, you're likely to need2 >>to restart it or reboot to clear a hung process. >>  E > Found the problem.  Note "f" had the listener process exit and the  N > AUDIT_SERVER was in RWMBX state.  The other two cluster member proceeded to & > place their processes in SUSP state. > G > I have no idea what went wrong, but I added a handler to clean up if   > something happens.  A This can happen if you have another mailbox declared as an audit  H listener, and the process (e.g third party monitoring software) reading  the mailbox falls over.   5 See the help for SET AUDIT/LISTENER for more details.    >   > Thanks for the response, Dave, >  > John   >  >    ------------------------------  % Date: Sat, 18 Dec 2004 17:45:31 -0500 = From: "John Gemignani, Jr." <john@nfw-invalid.cibtrikker.com> 9 Subject: Re: AUDIT_SERVER stops sending to the /LISTENER? 1 Message-ID: <PbqdnTmKntiTLlncRVn-pQ@adelphia.com>   8 "Paul Sture" <nospam@sture.homeip.net> wrote in message & news:32jg9cF3nl5m0U1@individual.net... > John Gemignani, Jr. wrote:C >> "David J Dachtera" <djesys.nospam@comcast.net> wrote in message  ( >> news:41C0E5D8.EB4FB074@comcast.net... >> >>>"John Gemignani, Jr." wrote:  >>> M >>>>For some reason today the audit server stopped sending sysuaf updates to   >>>>the M >>>>listener process.  I also no longer see the opcom alarms as well.  I did   >>>>not 7 >>>>make any changes to the system to make this happen.  >>>>( >>>>Here are the current audit settings: >>>> >>>>$ sho audit/all  >>>>List of audit journals: & >>>>  Journal name:           SECURITY4 >>>>  Journal owner:          (system audit journal)H >>>>  Destination:            DISK$LOGS:[SECURITY]SECURITY.AUDIT$JOURNAL% >>>>  Monitoring:             enabled E >>>>    Warning thresholds,   Block count:    100     Duration:    2   >>>> 00:00:00.0 E >>>>    Action thresholds,    Block count:     25     Duration:    0   >>>> 00:30:00.0  >>>>- >>>>Security auditing server characteristics: ! >>>>  Database version:       4.4 + >>>>  Backlog (total):        100, 200, 300 " >>>>  Backlog (process):      5, 2" >>>>  Server processing intervals:+ >>>>    Archive flush:        0 00:01:00.00 + >>>>    Journal flush:        0 00:05:00.00 + >>>>    Resource scan:        0 00:05:00.00 7 >>>>  Final resource action:  purge oldest audit events  >>>># >>>>Security archiving information: " >>>>  Archiving events:       none >>>>  Archive destination: >>>>1 >>>>System security alarms currently enabled for: 	 >>>>  ACL  >>>>  Authorization  >>>>  Audit:         illformed9 >>>>  Breakin:       dialup,local,remote,network,detached  >>>>  Logfailure: @ >>>>batch,dialup,local,remote,network,subprocess,detached,server >>>>1 >>>>System security audits currently enabled for: 	 >>>>  ACL  >>>>  Authorization  >>>>  Audit:         illformed9 >>>>  Breakin:       dialup,local,remote,network,detached J >>>>  Logfailure:    batch,dialup,local,remote,network,subprocess,detached >>>>% >>>>I don't get it.  Why did it stop?  >>> H >>>If you're not getting OPCOM messages, it's fair bet the OPCOM processI >>>has a problem. Depending what state OPCOM is in, you're likely to need 3 >>>to restart it or reboot to clear a hung process.  >>>  > F >> Found the problem.  Note "f" had the listener process exit and the L >> AUDIT_SERVER was in RWMBX state.  The other two cluster member proceeded * >> to place their processes in SUSP state. >>H >> I have no idea what went wrong, but I added a handler to clean up if  >> something happens.  > M > This can happen if you have another mailbox declared as an audit listener,  L > and the process (e.g third party monitoring software) reading the mailbox 
 > falls over.  > 7 > See the help for SET AUDIT/LISTENER for more details.  >  >>! >> Thanks for the response, Dave,  >> >> John    Paul,   H This is EXACTLY what happened.  The odd thing is that the victim system J didn't have any apparent problems (aside from the auditor stopping).  The 6 other two cluster members had processes in SUSP state.  K What tipped me off about the auditor is that SYSUAF changes were no longer  J being broadcast.  I was looking for problems in the systems with the SUSP L processes, not on the third party that was sitting quietly causing problems.   John     ------------------------------  # Date: Sat, 18 Dec 2004 20:57:00 GMT # From: Beach Runner <bob@nospam.com> * Subject: Re: DS10 on a NETGEAR DHCP Router< Message-ID: <wQ0xd.126683$8G4.86895@tornado.tampabay.rr.com>  2 I tried 4 cables. Bought a new cable. Worked fine.   Beach Runner wrote: F > A different thought. Is there a router that works with dhcp without ! > hitch on a 7.3-2 dchp, 100 full  >  >  > JF Mezei wrote:  >  >> Dave Froble wrote:  >>H >>>> I have a Netopia R9100 router which includes multiple RJ45 ports.   >>>> The router  >>>  >>> H >>> does not (easily) support dial-up access.  I also have a Multi-Tech  >>> router that J >>> supports dial-up as a backup link.  (I'm always on the backup link.)   >>> :-( E >>> So, how can I use the Multi-Tech as a dial-up access device, but  I >>> force all traffic through the Netopia?  Is this possible, given that   >>> each is connected  >> >> >> via >> >>> the integral hub/switch? >> >> >>K >> I assume that when you mention "dialup", you mean your site initiating a J >> dialup PPP connectioon to an ISP, or some user dialing into your site ? >>< >> dialup modem--[multitech]---*                +----[host1]1 >>                             |                | > >>                             |                +----[host2]  1 >> dslmodem--------------------+                | < >>                             |                +----[host3]1 >>                             |                | < >>                             +---[netopia]----+----[host4] >> >>I >> You then work on the netopia to have 2 possible configs: one where it   >> dealsL >> with your broadband provider, in which case, the netopia will forward all0 >> packets over to the modem's ethernet address. >>C >> Or you configure the netopia to forward all packets over to the   >> multitech, atF >> which poinmt the multitech acts as an ISP, gives your netopia some  >> other IP F >> address, acts as default gateway to your netopia and initiates the  >> dialup connection.  >>C >> Host1 through 4 don't see the difference since they all talk to   >> netopia only.J >> You need to ensure that DNS issues are dealt with (either the multitechI >> provides the dialup isp's DNS address upstream to netopia and netopid  	 >> to the J >> host1-4, or you hardcode some DSN server which is accessible from both 
 >> dsl and >> dialup connections. >> >>F >> The above config would require you put your hub on the WAN side of 
 >> netopia to G >> link netopia, dslmodem and mulltitech together, and you can use the   >> netopia'sD >> 4 port built=in switch to connect host1 through 4 to the netopia. >>G >> There are other ways to do this, such as putting both dsl and dialup K >> connections on the multitech, and then have the netopia only talk to the L >> multitech. However, this introduces some additional latency, since you're( >> going through 2 routers at all times. >>I >> Ideally, you need a more sophisticated router that can handle 2 links  	 >> to the I >> internet and automatically switch over to a second one when first one  	 >> fails.    ------------------------------  % Date: Sat, 18 Dec 2004 18:04:01 -0600 / From: Chris Scheers <chris@applied-synergy.com> $ Subject: Heirarchal network question, Message-ID: <32jur8F3m6ufpU1@individual.net>  F This is not really a VMS specific question, but I ask it here because ' there are some sharp network folk here.   F I have a client that has a small number of subnets.  There are 1 to 6 + machines in each subnet and 5 or 6 subnets.   I These subnets need to be isolated from each other, but they want each of  G the subnets to have Internet access.  Also, each subnet should be it's   own DHCP and NAT domain.  I I think I can do this with a heirarchy of broadband routers.  I envision  I connecting one router to the Internet connection, then a second layer of  F routers to the top level one, then the individual subnets each to its ? own second level router.  The routers can provide DHCP and NAT   functionality.  F Is there a way I can do this with a single router?  I've been looking F into VLANs, but the information I've seen so far doesn't say anything C about establishing seperate DHCP and NAT domains within the router.     Any advice would be appreciated.   Thanx!  G ----------------------------------------------------------------------- $ Chris Scheers, Applied Synergy, Inc.  B Voice: 817-237-3360            Internet: chris@applied-synergy.com    Fax: 817-237-3074   ------------------------------  % Date: Sat, 18 Dec 2004 22:32:36 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ( Subject: Re: Heirarchal network questionB Message-ID: <1103426613.1925c17269bbe2d725a491061f2d0516@teranews>   Chris Scheers wrote:  G > Is there a way I can do this with a single router?  I've been looking G > into VLANs, but the information I've seen so far doesn't say anything E > about establishing seperate DHCP and NAT domains within the router.   L VLANs are more an ethernet concept than a TCPIP concept.  Between VLAN-awareO switches, the ethernet packets are different and contain an additional payload.   I So if you can find a VLAN aware router that has DHCP capability, it could 1 theoretically discriminate between DHCP requests.   J Is there a reason why you need different subnets at the TCPIP level if you- provide ethernet blocks between the subnets ?    ------------------------------  % Date: Sat, 18 Dec 2004 11:19:16 -0800 * From: "Jack Peacock" <peacock@simconv.com> Subject: Re: HP exodus to Intel 2 Message-ID: <I42dnaBWSZUpH1ncRVn-pA@mpowercom.net>  9 "Rob Young" <young_r@encompasserve.org> wrote in message  - news:njSRwm13haOB@eisner.encompasserve.org... ? > You've know for "years" Jack that your VMS knowledge wouldn't A > be too valuable "in a few years."  But now you claim you're not : > looking forward to writing off those years.  You've been- > looking forward for quite some time , Jack.  > I I've been supporting NT/2000/2003 systems for about 8 years now, sitting  I alongside VMS servers.  I'm sure my story isn't unique among the readers  J here.  Given a preference I'd take VMS any day, but I like to eat and pay 7 bills.  So I cram in as much Microsoft as I can manage.   J Anyone who thinks VMS is still a career choice for the next twenty years, G well they might get lucky and find such a shop, but more likely it's a  H surefire ticket to eventual unemployment.  And it's not just VMS in the F downward spiral.  I wouldn't recommend SPARC machines for a long term I investment either (in fact I'm off to China this next week to debate the   point at a Sun installation).   K In the great scheme of things it has not been "for quite some time".  Look  L back over the last thirty something years and there have been any number of I similar transitions.  How many Algol programmers still around?  I'm sure  J there must be a few places still running some version of Univac Exec, but K when was the last time you saw a want ad for MCP, Kronos, or TOPS for that  M matter?  Anyone that's been in the business for a while knows that technical  H knowledge, especially the vendor-specific kind, has a useful but finite J half-life.  I'd sure like to have the time to earn more from the years of L VMS experience, hey, I'm as lazy as the next programmer, but it isn't going 
 to happen.    Jack Peacock    ------------------------------  % Date: Sat, 18 Dec 2004 17:10:24 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com>  Subject: Re: HP exodus to Intel , Message-ID: <41C4AACF.75697CDE@teksavvy.com>   icerq4a@spray.se wrote: @ > That is not correct. Itanium sells more than PaRisc and Alpha.  J HP revealed that their target this year was to grow IA64 revenus to 20% of; enterprise systems revenus. (I.E. it isn't even there yet).    ------------------------------    Date: 18 Dec 2004 21:23:21 -0600+ From: young_r@encompasserve.org (Rob Young)  Subject: Re: HP exodus to Intel 3 Message-ID: <gBb71H9xsl0F@eisner.encompasserve.org>   _ In article <I42dnaBWSZUpH1ncRVn-pA@mpowercom.net>, "Jack Peacock" <peacock@simconv.com> writes: ; > "Rob Young" <young_r@encompasserve.org> wrote in message    M > In the great scheme of things it has not been "for quite some time".  Look  N > back over the last thirty something years and there have been any number of @ > similar transitions.  How many Algol programmers still around?  B 	Poor example.  The correct question is how many Cobol programmersF 	are there.  Plenty.  So if there is still quite a Cobol infrstructure@ 	40 years out, what is your point?  Some languages are doomed to 	failure?  Sure.   >  I'm sure L > there must be a few places still running some version of Univac Exec, but M > when was the last time you saw a want ad for MCP, Kronos, or TOPS for that   > matter?     > 	Hard or impossible to get hardware even replacement parts for: 	many of what you mention there.  Would be kind of foolish) 	not to have a transition away from them.   F > Anyone that's been in the business for a while knows that technical J > knowledge, especially the vendor-specific kind, has a useful but finite  > half-life.    : 	Yeah... like the half-life of MVS and VMS.  Those are twoA 	good examples.  Surely Jack if you keep crowing about the future @ 	of VMS, someday it will come to pass it won't be around in some= 	distant future.  But you have to admit it is pretty funny to @ 	go back and read your posts 5 or so years ago talking about how: 	VMS is going away in 5 years, Compaq will kill it , etc.    				Rob  	    ------------------------------  % Date: Sun, 19 Dec 2004 01:32:24 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com>  Subject: Re: HP exodus to Intel B Message-ID: <1103437384.075b67df28fec0639932d8e56224ee9d@teranews>   Rob Young wrote:F >         distant future.  But you have to admit it is pretty funny toI >         go back and read your posts 5 or so years ago talking about how B >         VMS is going away in 5 years, Compaq will kill it , etc.  N Mr Young, I believe you are perfectly aware of how close VMS came to be killedO under Compaq's regime. So the above statement is rather ironic coming from you.   D Prediticting the demise of Linux when it is growing is not credible.  N But VMS has been on a decline, at least in appearance, for 15 years. VMS is noK longer the recipient of the biggest software library, a big sell for VMS in  the 1980s.    K Many potential customers wont even RFP for VMS solutions because they don't 9 want to invest in a platform that has been declared dead.   J VMS may not be officialy dead. But for all practical purposes, it has beenX relegated to such a small niche market that it has become irrelevant to the IT industry.  M The reason that there are on going discussions about VMS here is that many of ( us still believe that VMS has potential.  K But one by one, people come to realise that it is a losing battle not worth E fighting and the need to buy food and shelter forces all but the very G fortunate to prostitute themselves and work on Windows or Unix systems.    ------------------------------  % Date: Sat, 18 Dec 2004 23:53:45 +0100 * From: Paul Sture <nospam@sture.homeip.net>& Subject: Re: Interesting coding tidbit, Message-ID: <32jqnqF3mqdvfU1@individual.net>   Larry Kilgallen wrote:  h > In article <41C3E043.6070403@transgrid.com.au>, Paddy O'Brien <paddy.o'brien@transgrid.com.au> writes: >  >> >>Keith Cayemberg wrote: >> >>Lots snipped >> >>K >>>60's. The Unix and Windows designers didn't learn from history, and are   >>>now paying the price. >>>  >>D >>No, the users are paying the price.  BG is still making billions, 1 >>companies who use his crap are losing billions.  >  > @ > But since Windows is in a (large) niche market -- the desktop,C > the cost of what Bill hath wrought is intermixed with the general ? > cost of insufficiently centralized computing.  How many large > > buildings are cooled with individual room air conditioners ?  H A good point, as portayed by a company who found that they possessed no E less than 80 or so different models of printers, each with their own  4 stock of toner cartridges and maintenance contracts.  J OTOH, in practice many companies moved to decentralized computing because:  C a) the users got fed up with waiting for solutions from centralized      IT departments; = b) smaller computer solutions could be bought on departmental C     budgets, rather than going through central purchasing channels.    ------------------------------  % Date: Sun, 19 Dec 2004 00:07:42 +0100 * From: Paul Sture <nospam@sture.homeip.net> Subject: Re: soon to be gone, Message-ID: <32jrhvF3ld7peU1@individual.net>   John Smith wrote:    > sethmuce@retronet.net wrote: > $ >>John Smith <a@nonymous.com> wrote: >> >>>>   You can't go back.  >>>  >>> / >>>Here's another building they should implode: % >>>3000 HANOVER STREET, PALO ALTO, CA  >>F >>Hey, I work right across the street from them -- so could you please* >>give me a little warning first?  Thanks. >  >  > J > Implode, not explode.....mostly due to the vacuous conditions inside the > building.  >  > @ Here's an example. The nearest building was only 17 metres away:  * http://www.implosionworld.com/bradford.htm   http://tinyurl.com/4vugd  F They ended up replacing someone's window and repairing another bit of  superficial damage.    ------------------------------    Date: 18 Dec 2004 14:48:17 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) $ Subject: Re: Time to revive Emerald?3 Message-ID: <LsFqZkQneYG7@eisner.encompasserve.org>   f In article <jFWwd.20683$pb.1392287@news20.bellglobal.com>, "Neil Rieck" <n.rieck@sympatico.ca> writes: > 6 > "Dave Froble" <davef@tsoft-inc.com> wrote in message( > news:41C37973.5020800@tsoft-inc.com... >> Neil Rieck wrote: >> > [...snip...] >>J >> There are some who would view this as a disadvantage, not an advantage.B >> Regardless of whatever target language, rewriting 1/3 of VMS is2 >> non-trivial and many errors will be introduced. >> > J > At a 2003 OpenVMS seminar in Ottawa, an HP employee stated the followingI > fact: "Porting VMS from VAX to Alpha required touching 1200 s/w modules A > while porting from Alpha to Itanium required touching only 200"  > L > I'm not sure what is meant by "touch", or whether this employee was trying  D What is more important is what is meant by "not touch".  If there isB no need to change a module in any way that drastically reduced the effort to validate changes.   N > to trivialize the porting effort. So you may be right in stating that 1/3 ofG > VMS needed a rewrite. On the flip side, I know that the port to Alpha H > involved a "code cleanup" and they probably added lexicals to simplifyL > future porting operations. That said, touching 1/6 of the OS doesn't sound, > as bad (or as expensive) as rewriting 1/3.  > I believe the 1/3 figure was to convert lots of modules to theA poster's favorite language, not to accomplish a business purpose.   D Looking at the VMS V7.3-2 source listings kit I see about 10137 .LIS
 files, so:   	1200 modules = 1/8  and  	200 modules = 1/50   A so the porting effort (which did _not_ include a total rewrite of B every module touched) was much less invasive than the 1/3 proposed$ for rewrite to adopt a fad language.   ------------------------------  % Date: Sat, 18 Dec 2004 17:07:35 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> $ Subject: Re: Time to revive Emerald?, Message-ID: <41C4AA26.E0448CEE@teksavvy.com>   Neil Rieck wrote: H > If the renovated Alpha code base was designed to only deal with 64-bitC > CPU's, then using it to port to IA-32 isn't going to be possible.   + Forget about porting to 32 or 16 bit 8086s.   L The "current" 8086 is 64 bits. There is no point in porting VMS to the olderG versions of the 8086. So, porting VMS to the 64 bit version of the 8086 E shouldn't be that difficult, provided the chip provides the necessary 5 functions, and from what has been said here, it does.    ------------------------------  % Date: Sat, 18 Dec 2004 16:31:10 -0800 # From: "Tom Linden" <tom@kednos.com> $ Subject: Re: Time to revive Emerald?( Message-ID: <opsi8bp8d6zgicya@hyrrokkin>  8 On 18 Dec 2004 08:07:20 -0800, <icerq4a@spray.se> wrote:  & > As a normal UNIX, except clustering,> > Tru64 was not in any way better than the other UNIX systems.  1 I might dispute that, but, the what is the point.    --  C Using Opera's revolutionary e-mail client: http://www.opera.com/m2/    ------------------------------  % Date: Sat, 18 Dec 2004 13:01:02 -0500 4 From: David R. Beatty <QWDavidER.TYBeattyUI@sas.com>5 Subject: Re: Vanquishing the Smoke and Fog with Facts 8 Message-ID: <h0s8s09ng13af9oors2hpg7a3jdjttgkes@4ax.com>  @ On 17 Dec 2004 16:19:36 -0800, "mas" <mas769@hotmail.com> wrote:  + >This maybe of, ahem, interest to you guys.  > j >http://www.realworldtech.com/forums/index.cfm?action=detail&PostNum=2961&Thread=6&entryID=43332&roomID=11 > j >http://www.realworldtech.com/forums/index.cfm?action=detail&PostNum=2961&Thread=8&entryID=43337&roomID=11 > k >http://www.realworldtech.com/forums/index.cfm?action=detail&PostNum=2961&Thread=10&entryID=43346&roomID=11  >  >  >  >> John Smith wrote: >>	 >> [....]  >>H >> Does this mean that in 2008 the axe will fall on Itanic if  purchases >of the H >> chips isn't sufficient to interest Intel in continuing production and >> further development?  >>H >> How many times does the OpenVMS Engineering team have to drop all the >newD >> features and improvements they ought to be working on in order to >port to >> the 'chip of the week'? >>' >> I guess the answer to that would be: = >> 1) No advertising/marketing of OpenVMS = no new customers. - >> 2) No new customers = continued attrition. 4 >> 3) Continued attrition = EOL announcement for VMSF >> 4) EOL announcement = no port to Power ot Opteron or anything else.G >> 5) No port = all us old dogs have to learn new tricks in order to be  >able to >> retire comfortably. >> >>? >> Seems to me that Alpha is the way to go....HP still owns the 
 >intellectual H >> property rights, correct?  Oh, I forgot...HP doesn't do its own thingB >> anymore....they just Invent new ways to run a business into the >ground.  2 Mas, that is probably the same guy that wrote this article:  ) http://www.theinquirer.net/?article=20242   = John Smith referred to it in the post "The Other Shoe Drops".    David Beatty   ------------------------------  % Date: Sat, 18 Dec 2004 16:27:20 -0500 2 From: Chip Coldwell <coldwell@physics.harvard.edu>$ Subject: Re: VESA/VGA BIOS emulationA Message-ID: <Pine.LNX.4.61.0412181547060.23971@frank.harvard.edu>   ) On Thu, 16 Dec 2004, Chip Coldwell wrote:    > E > The Alpha SRM console supports a PC BIOS emulation callback routine G > (Alpha Architecture Reference Manual, 3rd Ed, part III sect 2.3.7.3). D > I've been using this function from VMS to set VGA video modes on aF > video card not currently supported by OpenVMS Alpha, and since I can> > see the results on screen I know this part of my code works. > B > One of the arguments the BIOS_EMUL function takes is an array ofF > longwords that correspond to the emulated Intel x86 registers.  SomeF > of the VESA extended BIOS functions (VBE 3.0) take parameters passedE > by reference in the x86 registers, with the pointers in ES:DI using E > the godawful Intel 20-bit real-mode segmented addressing scheme.[1]  > @ > My question is: how do I map from an OS virtual address to the. > values in the emulated 8086 ES:DI registers? > G > For example, if I put the value 0x4F00 into the emulated AX register, F > and a pointer to a 512 byte buffer in the ES:DI registers, then call= > int 0x10, the VESA BIOS will fill in a bunch of very useful H > information in the 512 byte buffer[2].  I can do all of this using theF > SRM BIOS emulator callback, and I know it does something because theG > AX register has the value 0x004F on return (byte-swap of the input AX 4 > value), which the documents say indicates success. > C > I suspect there is a region of virtual address space set aside to H > serve as memory for the x86 emulator, and that addresses I pass in theE > ES:DI registers turn into offsets from the base of that region.  If H > that's true, then what I really need to know is what that base address > is.  > C > The BIOS_EMUL console callback can be passed an "additional_data" I > parameter in the r20 register, but putting the virtual address in there K > didn't work.  P0 space includes virtual addresses that fit in 20 bits, so L > I tried stuffing one into the x86 segment/offset registers but that didn't7 > work either.  I even tried doing both simultaneously.  > D > I noticed that XFree86 solves this problem by having their own x86F > emulator that runs in the OS (see the contents of "xc/extras/x86emu"F > in their source distribution).  This enables things like the XFree86F > VESA driver to work on Alpha.  I could do something similar; I couldB > even use the emulator that comes with XFree86 since it was thereF > before their license changed, but it seems a shameful duplication of5 > effort if the SRM console already does what I need.  > < > I can post my code if anyone thinks that would be helpful. >  > Notes: > L > [1] The 8086 segment/offset scheme represents a 20-bit address by a 16-bitJ > segment register and a 16-bit offset register where the physical address@ > is calculated as (segment << 4) + offset.  Yes, there are manyC > segment/offset combinations that yeild the same physical address.  > 3 > [2] See http://www.vesa.org/vbe3.pdf for details.  >  > Chip >  >   B Here I am following up to my own post, and cross-posting to boot. I Netiquette goes out the window.  But maybe getting this information into  3 the Google archive will help somebody else someday.   A While reading the Book of Ruth (OpenVMS Alpha Internals and Data  H Structures: Memory Management, by Ruth E. Goldenberg) I stumbled across ( the following tantalizing quote on p. 78  F     "At boot time, the console determines the memory configuration andI      passes its findings as well as its own memory use to OpenVMS through A      a data structure called the hardware restart parameter block F      (HWRPB). ... Chapter Boostrap Processing provides more details on*      the HWRPB and associated structures."  D I already knew about the HWRPB, of course, but the phrase "its [the D console's] *OWN* memory use" was interesting.  The SRM console BIOS K emulator must stash the results of these BIOS calls somewhere in the range  I of "its own memory use", so all I need to do is search through the range  H of addresses set aside for the console for a block of memory that looks  like what I want.   H Fortunately, the VESA standard specifies that the block returned by the I 0x4f00 BIOS call starts with the ASCII-encoded string 'VESA' followed by  K the version number 0x0300 (see p.25 of the vbe3.pdf document cited above).  # That gave me something to look for.   J Furthermore, the console routine block (CRB) pointed to by HWRPB contains J a virtual-physical map section.  The AARM describes the format in detail, J but doesn't give any hint as to the purpose of these mappings.  I took an F educated guess that the emulated x86 memory used by the BIOS_EMUL SRM J callback must be located in one of these mappings.  The following program J searches through the CRB mappings for the occurrance of the string 'VESA' # and dumps some memory following it:    #include <hwrpbdef.h>  #include <starlet.h> #include <stdio.h>   static char buf[256];   . static const char *crb_search(const char *str) { !      extern HWRPB *exe$gpq_hwrpb;       HWRPB_CRB *crb;      VAPAMAP *map;      const char *ptr;       long long i, len;  &      for(len = 0; *str; str++, len++);      str -= len;  4      crb = (HWRPB_CRB *) ((long long)exe$gpq_hwrpb +&  	exe$gpq_hwrpb->hwrpb$iq_crb_offset);  3      map = (VAPAMAP *) &crb->hwrpb_crb$iq_vapa_map;   3      for(i=0; i<crb->hwrpb_crb$iq_map_count; i++) { 7  	long long size = map[i].hwrpb_vapamap$iq_page_count * '  	    exe$gpq_hwrpb->hwrpb$iq_pagesize; 0  	for(ptr = (char *) map[i].hwrpb_vapamap$iq_va;C  	    ptr < (char *) map[i].hwrpb_vapamap$iq_va + size - len + 1LL;   	    ptr++) {  		const char *p, *s; 3  		for(p = ptr, s = str; *s && *s == *p; p++, s++); 
  		if(!*s)  		    goto found;  	}       }        return 0; found:      for(i=0;i<sizeof(buf);i++)   	buf[i] = ptr[i];      return ptr; }   
 int main() { "      int arglist[2], i, *ptr, adr;        arglist[0] = 1;      arglist[1] = (int) "VESA"; +      adr = sys$cmkrnl(crb_search, arglist);       if(adr) {  	ptr = (int *) buf;  	for(i=0; i<8; i++).  	    printf("%08x: %08x\n", adr+4*i, ptr[i]);      }        return 0; }    and here's the output:   $ run crb_search 8b8622f5: 41534556 8b8622f9: 00000300 8b8622fd: 00010000 8b862301: 00000000 8b862305: 00000000 8b862309: 00000000 8b86230d: 00000000 8b862311: 00000000  G 0x41534556 is 'VESA' in ASCII on a little-endian machine.  The version  I code 0x0300 occupies the next two bytes.  I'm pretty sure I've found it.  H But there are still some strange things about it, such as the fact that  it's not longword-aligned.  & The reverse-engineering continues ....   Chip   --   Charles M. "Chip" Coldwell System Administrator Harvard Physics Department 617-495-3388   ------------------------------  % Date: Sat, 18 Dec 2004 22:50:27 +0100R< From: =?iso-8859-1?q?M=E5ns_Rullg=E5rd?= <mru@inprovide.com>$ Subject: Re: VESA/VGA BIOS emulation1 Message-ID: <yw1x8y7v6zto.fsf@ford.inprovide.com>e  4 Chip Coldwell <coldwell@physics.harvard.edu> writes:   [reverse engineering]e  H > 0x41534556 is 'VESA' in ASCII on a little-endian machine.  The versionF > code 0x0300 occupies the next two bytes.  I'm pretty sure I've foundH > it. But there are still some strange things about it, such as the fact! > that it's not longword-aligned.l  C That does sound somewhat odd.  Does the following data seem to make- sense?  ( > The reverse-engineering continues ....  @ Please keep us updated.  The information may be useful some day.   -- e
 Mns Rullgrdh mru@inprovide.com:   ------------------------------  # Date: Sat, 18 Dec 2004 20:05:05 GMTnG From: "Richard S. Shuford" <shuford@list.stratagy.REM0VE-THlS-PART.com>"$ Subject: Re: VMS with BULL terminals. Message-ID: <rshu_20041218200505@stratagy.com>  2 Volker N. Englisch <englisch.ahr(at)gmx.de> wrote: |eG | I have several "Bull Questar 310" attached to a unix machine in VT220>E | emulation mode. They work very fine. Now I want to connect to a VMSi | host using telnet... | G | After calling an editor on the VMS machine that uses fullscreen mode,iC | there is only line 24 remaining for any action on the screen. The>H | output of the host scrolls only in this single line. The terminal line, | on the VMS host is set to VT200 (-series). |SG | One DEC VT510 in VT220 emulation mode, working as the system console,a | doesn't have these problems.     Volker:t  @ Your problem statement does not contain enough precise detail to0 allow anybody to be certain of what is going on.  A It would be helpful, for instance, to know if "an editor" is EVE, , or EDT, or some vi clone, or something else.  ? Do you mean that your terminal is connected by an RS-232 serial = cable to the Unix host, and then you are running the "telnet"0< command on the Unix host to connect through to the VMS host?7 And that the Bull terminal works OK with the Unix host?   > The following statement is unclear:  "The terminal line on the@ VMS host is set to VT200 (-series).  Do you mean that the output< of "SHOW TERMINAL" indicates a "VT200-series" terminal type?  > The following statement is unclear:  "...there is only line 24A remaining for any action on the screen."  Do you mean that all ofa= the computer's output gets rendered on the bottom line of thed display?  ; It is possible that the Bull Questar 310 does not correctlyL? emulate the DEC VT220 in some respect.  The "scrolling regions"nA feature would be under suspicion, but it could be something else.a> You could run the "vttest" program on the Unix machine to test5 the display-feature compatibility of the Questar 310.e  2     http://invisible-island.net/vttest/vttest.html  ? The EDT and EVE editors infamously exploited tricky features oft@ the DEC terminals, so incompatibility often is revealed by them.  @ Another problem might be lack of 8-bit channel transparency. YouA could maybe force the VMS machine to send 7-bit control sequencesr by some command like       SET TERMINAL/NOEIGHTBIT.  >     (No longer having access to a VMS machine, I'm not certainE     that's the correct syntax.  Use the "HELP" system to look it up.)e  ; Or you could throttle back the VMS machine's notion of whatg+ terminal type it's talking to, by trying...h       SET TERMINAL/TYPE=VT102h   Hope that helps.       ...RSSo   -- i' Your cow joke might be worth a Frisbee.v? http://www.stonyfield.com/weblogarchives/DailyScoop/000651.htmle   ------------------------------   End of INFO-VAX 2004.702 ************************