Re: TCP/IP relationship to OMVS
I've been playing around with COBOL and the Unix callable services. Kinda cool! Glad they are not limited to assembler. From: IBM Mainframe Discussion List on behalf of David Crayford Sent: Friday, February 5, 2021 12:24 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: TCP/IP relationship to OMVS On 5/02/2021 1:32 am, Charles Mills wrote: > TCP/IP certainly makes use of lots of UNIX services. Yes, and it goes both ways. I'm working a project right now with one of the original IBM OMVS developers and he told me that the BPX callable services wrap Comms Server APIs and then the LE C/C++ runtime wrap the BPX services. AF_UNIX sockets are handled in OMVS. It was an inspired decision by IBM to provide callable services for z/OS UNIX and not another macro interface like EZASMI. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
That depends on what you mean by recent. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 From: IBM Mainframe Discussion List [IBM-MAIN@LISTSERV.UA.EDU] on behalf of Paul Gilmartin [000433f07816-dmarc-requ...@listserv.ua.edu] Sent: Thursday, February 4, 2021 2:37 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: TCP/IP relationship to OMVS On Thu, 4 Feb 2021 09:32:58 -0800, Charles Mills wrote: > >I believe TCP/IP and several of its "children" such as FTP server run as UNIX >daemons. > Some such children are recent adoptees. I believe FTP on MVS far antedates OMVS. (But was that an ISV offering? IIRC an Intel Fastpath(?) that masqueraded as a CTC.) -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
There was also an ISV TCP/IP implementation from Interlink. I recall at the time modifying Lionel's excellent XMITIP to use Interlink TCP. Dana On Thu, 4 Feb 2021 23:46:40 +, Frank Swarbrick wrote: >Looks like OS/390 V2R4 (Sept 1997?) contained "A new TCP/IP stack, for >applications using OS/390 UNIX System Services (formerly called OpenEdition) >sockets". > >Prior to that was TCP/IP Version 3 Release 2, which I believe was based on a >port of TCP/IP for VM (and written in Pascal!). Or something like that. > -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
On 5/02/2021 1:32 am, Charles Mills wrote: TCP/IP certainly makes use of lots of UNIX services. Yes, and it goes both ways. I'm working a project right now with one of the original IBM OMVS developers and he told me that the BPX callable services wrap Comms Server APIs and then the LE C/C++ runtime wrap the BPX services. AF_UNIX sockets are handled in OMVS. It was an inspired decision by IBM to provide callable services for z/OS UNIX and not another macro interface like EZASMI. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
On Thu, 4 Feb 2021 17:10:55 -0800, Charles Mills wrote: >Are there not occasional fleeting references here and there to "the Pascal >TCP stack"? > That had the notorious GIVESOCKET/TAKESOCKET, needed because prior to OMVS, descriptor inheritance was unheard of. Orthogonal concepts. I don't know whether inheritance or sharing among siblings is the more powerful feature. >The first I personally wrote any z/OS socket code was around 2010. (I had an >employee who wrote socket code for Z in the 90's but I am not the least bit >intimate any more with the details. I seem to recall that perhaps it was a >SAS-licensed TCP stack? Does that ring a bell?) > IBM implemented SPFGUI(?) (ancestor of WSA?) in SAS/C because at that time SAS/C had better TCP/IP support than IBM's product. They didn't much publicize it. That caused us problems because we were trying to market a product written in SAS/C, not networked at all. But we had conflicts with incompatible service levels of the runtime library (DLL?) between our product and IBM's. At one time, the only descriptors supported by SAS/C were sockets. RXSOCKET had (has?) optional ASCII<->EBCDIC translation Does that still exist? https://www.ibm.com/support/knowledgecenter/SSB27U_6.4.0/com.ibm.zvm.v640.dmsb1/rxstips.htm -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
VMCF and TNF may still be on your systems for the few remnants depending on the PASCAL stack API. It was ported out of VM and did not perform well either. Another factor which slowed the adoption of TCP/IP in the z/OS (them MVS world) was the IBM Communications group resistance. They had all their bets and plans revolving around keeping and protecting SNA > -Original Message- > From: IBM Mainframe Discussion List On > Behalf Of Charles Mills > Sent: Thursday, February 04, 2021 5:11 PM > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: TCP/IP relationship to OMVS > > Are there not occasional fleeting references here and there to "the Pascal > TCP stack"? > > The first I personally wrote any z/OS socket code was around 2010. (I had an > employee who wrote socket code for Z in the 90's but I am not the least bit > intimate any more with the details. I seem to recall that perhaps it was a > SAS-licensed TCP stack? Does that ring a bell?) > > Anyway, the Comm Server doc at that time (2010) -- I see it still around for > z/OS V1R13 -- had a manual "IP Sockets Application Programming Interface > Guide and Reference." I think I originally got going in 2010 using that doc, > and it was not right at all for the TCP implementation I was using, which is > correctly documented in the C Library manual. I think it may be doc for the > Pascal stack. Does that make sense? I quickly got straightened out. > > Charles > > > -Original Message- > From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] > On > Behalf Of Frank Swarbrick > Sent: Thursday, February 4, 2021 3:47 PM > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: TCP/IP relationship to OMVS > > Looks like OS/390 V2R4 (Sept 1997?) contained "A new TCP/IP stack, for > applications using OS/390 UNIX System Services (formerly called > OpenEdition) > sockets". > > Prior to that was TCP/IP Version 3 Release 2, which I believe was based on a > port of TCP/IP for VM (and written in Pascal!). Or something like that. > > References: > https://urldefense.com/v3/__https://www- > 01.ibm.com/common/ssi/ShowDoc.wss?docURL=*common*ssi*rep_ca*6*8 > 7__;Ly8vLy8!!JmPEgBY0HMszNaDT!_11xmxKMzbWFPba3c4Ph4oLlxM8bXcTY > A6cYdNqZeSdE15KOYRjphExehsA6FA$ > 7/ENUSZP97-0486/index.html&request_locale=en > https://urldefense.com/v3/__https://www- > 01.ibm.com/common/ssi/rep_ca/9/897/ENUS296- > 339/index.html__;!!JmPEgBY0HMszNaDT!_11xmxKMzbWFPba3c4Ph4oLlxM8 > bXcTYA6cYdNqZeSdE15KOYRjphEzIZ2CIDg$ > https://urldefense.com/v3/__https://groups.google.com/g/bit.listserv.ibm- > main/c/M7wIMma2Ig4__;!!JmPEgBY0HMszNaDT!_11xmxKMzbWFPba3c4Ph > 4oLlxM8bXcTYA6cYdNqZeSdE15KOYRjphEyixtv9-w$ > > > ________ > From: IBM Mainframe Discussion List on > behalf of > Pew, Curtis G > Sent: Thursday, February 4, 2021 4:09 PM > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: TCP/IP relationship to OMVS > > On Feb 4, 2021, at 4:24 PM, Mike Schwab > wrote: > > > > z/OS FTP required a RACF OMVS flag in 2001 when I started submitting > > dumps to IBM. > > > > I have a vague recollection that in the early 1990s IBM provided a TCP/IP > stack on MVS that was ported from VM, but sometime in the OS/390 era it > was > replaced by a new Unix-based implementation. > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
Are there not occasional fleeting references here and there to "the Pascal TCP stack"? The first I personally wrote any z/OS socket code was around 2010. (I had an employee who wrote socket code for Z in the 90's but I am not the least bit intimate any more with the details. I seem to recall that perhaps it was a SAS-licensed TCP stack? Does that ring a bell?) Anyway, the Comm Server doc at that time (2010) -- I see it still around for z/OS V1R13 -- had a manual "IP Sockets Application Programming Interface Guide and Reference." I think I originally got going in 2010 using that doc, and it was not right at all for the TCP implementation I was using, which is correctly documented in the C Library manual. I think it may be doc for the Pascal stack. Does that make sense? I quickly got straightened out. Charles -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Frank Swarbrick Sent: Thursday, February 4, 2021 3:47 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: TCP/IP relationship to OMVS Looks like OS/390 V2R4 (Sept 1997?) contained "A new TCP/IP stack, for applications using OS/390 UNIX System Services (formerly called OpenEdition) sockets". Prior to that was TCP/IP Version 3 Release 2, which I believe was based on a port of TCP/IP for VM (and written in Pascal!). Or something like that. References: https://www-01.ibm.com/common/ssi/ShowDoc.wss?docURL=/common/ssi/rep_ca/6/87 7/ENUSZP97-0486/index.html&request_locale=en https://www-01.ibm.com/common/ssi/rep_ca/9/897/ENUS296-339/index.html https://groups.google.com/g/bit.listserv.ibm-main/c/M7wIMma2Ig4 From: IBM Mainframe Discussion List on behalf of Pew, Curtis G Sent: Thursday, February 4, 2021 4:09 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: TCP/IP relationship to OMVS On Feb 4, 2021, at 4:24 PM, Mike Schwab wrote: > > z/OS FTP required a RACF OMVS flag in 2001 when I started submitting > dumps to IBM. > I have a vague recollection that in the early 1990s IBM provided a TCP/IP stack on MVS that was ported from VM, but sometime in the OS/390 era it was replaced by a new Unix-based implementation. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
Looks like OS/390 V2R4 (Sept 1997?) contained "A new TCP/IP stack, for applications using OS/390 UNIX System Services (formerly called OpenEdition) sockets". Prior to that was TCP/IP Version 3 Release 2, which I believe was based on a port of TCP/IP for VM (and written in Pascal!). Or something like that. References: https://www-01.ibm.com/common/ssi/ShowDoc.wss?docURL=/common/ssi/rep_ca/6/877/ENUSZP97-0486/index.html&request_locale=en https://www-01.ibm.com/common/ssi/rep_ca/9/897/ENUS296-339/index.html https://groups.google.com/g/bit.listserv.ibm-main/c/M7wIMma2Ig4 From: IBM Mainframe Discussion List on behalf of Pew, Curtis G Sent: Thursday, February 4, 2021 4:09 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: TCP/IP relationship to OMVS On Feb 4, 2021, at 4:24 PM, Mike Schwab wrote: > > z/OS FTP required a RACF OMVS flag in 2001 when I started submitting > dumps to IBM. > I have a vague recollection that in the early 1990s IBM provided a TCP/IP stack on MVS that was ported from VM, but sometime in the OS/390 era it was replaced by a new Unix-based implementation. -- Pew, Curtis G curtis@austin.utexas.edu -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
I installed the U of Wisconsin DARPA project TCP/IP VM/CMS stack around 1986-ish, and (as I remember) it was indeed ported to MVS. IBM also rolled out their own VM/CMS stack version, and we had quite a jolly time converting. UW code had enhancements that IBM didn't have yet, and vice versa. The network connection was through a DACU (which featured an IBM PC running it) to a Network Systems Corp Hyperchannel Wisely, IBM rewrote the original Pascal code, but it was fun and educational going though the source while we had it. David -Original Message- From: Pew, Curtis G To: IBM-MAIN@LISTSERV.UA.EDU Sent: Thu, Feb 4, 2021 6:09 pm Subject: Re: TCP/IP relationship to OMVS On Feb 4, 2021, at 4:24 PM, Mike Schwab wrote: > > z/OS FTP required a RACF OMVS flag in 2001 when I started submitting > dumps to IBM. > I have a vague recollection that in the early 1990s IBM provided a TCP/IP stack on MVS that was ported from VM, but sometime in the OS/390 era it was replaced by a new Unix-based implementation. -- Pew, Curtis G curtis@austin.utexas.edu -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
On Feb 4, 2021, at 4:24 PM, Mike Schwab wrote: > > z/OS FTP required a RACF OMVS flag in 2001 when I started submitting > dumps to IBM. > I have a vague recollection that in the early 1990s IBM provided a TCP/IP stack on MVS that was ported from VM, but sometime in the OS/390 era it was replaced by a new Unix-based implementation. -- Pew, Curtis G curtis@austin.utexas.edu -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
z/OS FTP required a RACF OMVS flag in 2001 when I started submitting dumps to IBM. On Thu, Feb 4, 2021 at 1:37 PM Paul Gilmartin <000433f07816-dmarc-requ...@listserv.ua.edu> wrote: > > On Thu, 4 Feb 2021 09:32:58 -0800, Charles Mills wrote: > > > >I believe TCP/IP and several of its "children" such as FTP server run as > >UNIX daemons. > > > Some such children are recent adoptees. I believe FTP on MVS far antedates > OMVS. (But was that an ISV offering? IIRC an Intel Fastpath(?) that > masqueraded as a CTC.) > > -- gil > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
tcp/ip on MVS far antedates OMVS... TCP/IP (FAL) was around in the MVS/ESA days (and OS/390 I know for sure). Joe On Thu, Feb 4, 2021 at 1:37 PM Paul Gilmartin < 000433f07816-dmarc-requ...@listserv.ua.edu> wrote: > On Thu, 4 Feb 2021 09:32:58 -0800, Charles Mills wrote: > > > >I believe TCP/IP and several of its "children" such as FTP server run as > UNIX daemons. > > > Some such children are recent adoptees. I believe FTP on MVS far antedates > OMVS. (But was that an ISV offering? IIRC an Intel Fastpath(?) that > masqueraded as a CTC.) > > -- gil > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
On Thu, 4 Feb 2021 09:32:58 -0800, Charles Mills wrote: > >I believe TCP/IP and several of its "children" such as FTP server run as UNIX >daemons. > Some such children are recent adoptees. I believe FTP on MVS far antedates OMVS. (But was that an ISV offering? IIRC an Intel Fastpath(?) that masqueraded as a CTC.) -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
On Thu, 4 Feb 2021 at 12:25, Joe DeChirico wrote: > Is there any information available on the relationship between TCP/IP and > OMVS? I'm not sure if I (or others) fully understand your question. But one way of looking at it is to say that any TCP/IP implementation is really just a Physical File System (PFS) - AKA Installable File System - that plugs into the back end of the more general UNIX file system scheme. This is described in extreme detail in the book z/OS UNIX System Services File System Interface Reference, which you can find as a PDF or in the Knowledge Center. This diagram (called Figure 1) https://www.ibm.com/support/knowledgecenter/SSLTBW_2.4.0/com.ibm.zos.v2r4.bpxb500/bpx1fr_System_structure.htm provides a good overview. A TCP/IP implementation is one of the Physical File Systems (actually a Physical Socket System, but they don't use that phrase) in the lower left corner. For purposes of this diagram, any program running on z/OS that uses TCP/IP is in the box in the top left called "z/OS UNIX Programs". There can be as many PFSs as you like, and they can run either in the UNIX kernel or in a separate address space. Of course it's not "normal" to write your own physical file or socket system, but it can be done. Tony H. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
The current TCP/IP is a Unnix application, which means that it must be dubbed. A Unix Zystem Services application can use the legacy MVS services. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 From: IBM Mainframe Discussion List [IBM-MAIN@LISTSERV.UA.EDU] on behalf of Joe DeChirico [joe.dechir...@csi-international.com] Sent: Thursday, February 4, 2021 12:14 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: TCP/IP relationship to OMVS Hi Is there any information available on the relationship between TCP/IP and OMVS? Thanks Joe DeChirico -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: TCP/IP relationship to OMVS
What specifically? I believe TCP/IP and several of its "children" such as FTP server run as UNIX daemons. There is not so much of a classic MVS versus OMVS dichotomy as one might imagine. USS is a fundamental part of z/OS. It's not so much this "thing" that coexists with MVS as a set of services, a set of APIs, a way to utilize MVS services. Almost any program -- any conventional, batch, run from EXEC PGM=, get its input from DDs program -- can and may utilize one or more UNIX services. TCP/IP certainly makes use of lots of UNIX services. Charles -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Joe DeChirico Sent: Thursday, February 4, 2021 9:15 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: TCP/IP relationship to OMVS Hi Is there any information available on the relationship between TCP/IP and OMVS? Thanks Joe DeChirico -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
TCP/IP relationship to OMVS
Hi Is there any information available on the relationship between TCP/IP and OMVS? Thanks Joe DeChirico -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN