jeff%colgate.csnet@RELAY.CS.NET (Jeff Norden) (01/11/88)
Is there, is there, balm of Gilead? I have a feeling that the answer will be "Nevermore", but... . I have been trying to get three local unix machines to send mail and files back and forth with uucp. The machines are connected via terminal servers - which act pretty much like a terminal switch. The problem is that the servers are too smart for their own good, and they try to do flow control, autobaud, auto-parity, etc. A major stumbling block has been the complete lack of documentation on how the internals of uucp work - ie what exactly is the "g" protocol, and which ascii characters does uucp actually use? What I am looking for is something akin to the manual that I know exists kermit. I have access to the source for uucp, but neither the time nor desire to start wading through it. Is the lack of a generally distributed explanation of the protocol considered essential to security on uucp? It seems like a security system which is based on the assumption that people will be too lazy to read the source code is a pretty poor system - it won't foil any truly nasty folks, and its a real pain for everybody else. Anyway, I would really appreciate it if someone would mail me a copy of such a manual if it exists. Lacking that, I've included below some bits of debugging output which I am trying to comprehend. Any hints will be appreciated. Many thanks in advance. -Jeff Norden jeff@colgate.csnet Dept of Mathematics, Colgate University Hamilton, NY 13346. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ This is from a vax 11/750 running bsd4.3 % uucico -r1 -x999 -scolgmath uucp colgmath (1/9-22:34-12173) DEBUG (Local Enabled) finds (colgmath) called ***[uucico succsessfully calls up the machine colgmath and logs in] uucp colgmath (1/9-22:34-12173) SUCCEEDED (call to colgmath ) imsg looking for SYNC< nuucp UNIX System V Release 3.1 AT&T 3B2 colgmath Copyright (c) 1984 AT&T All Rights Reserved \20> imsg input<Shere=colgmath\0 ***[stuff which I gather is initial handshaking and protocol startup] Proto started g protocol g uucp colgmath (1/9-22:34-12173) OK (startup tty03 9600 baud) *** TOP *** - role=MASTER gnamef returns . bldflst rejects . gnamef returns .. bldflst rejects .. gnamef returns C.colgmatnZsd0 bldflst returns 1 jeff colgmath (1/9-22:34-12173) REQUEST (S /a/jeff/junk ~/ jeff) expfile type - 1, wrktype - S wmesg 'S' /a/jeff/junk ~/ jeff -dc D.0 644 send 0210 rmesg - 'S' rec h->cntl 041 state - 010 rec h->cntl 0211 PKCGET stall for 0.046800 sec pkcget: alarm 4001 rec h->cntl 0211 PKCGET stall for 0.023900 sec pkcget: alarm 7002 ***[a series of messages simialar the last three] state - 06000 got FAIL jeff colgmath (1/9-22:36-12173) BAD READ (expected 'S' got FAIL (2)) cntrl - -1 jeff colgmath (1/9-22:36-12173) FAILED (conversation complete) In particular, what do the "rec h->cntl xxxx" and "pkcget: alarm xxxx" messages mean?
mjh@uunet.UU.NET (Mark J. Hewitt) (01/13/88)
As far as I know, there is no UUCP protocol manual because the protocols
are proprietary to AT&T. The pckget: alarm xxxx are read timeouts. I
suspect you'll have to read the source, or find someone who knows it
well. I recall someone's e-mail footer containing the edict: "Use the
force, read the source".
Mark J. Hewitt
usenet: ...!{mcvax,uunet,ukc}!kernel!mjh
other: mjh%kernel.uucp@ukc.ac.uk
voice: (+44) 532 465311
fax: (+44) 532 420183
paper: Kernel Technology Ltd, 21 Queen Street,
Leeds, LS1 2TW, West Yorkshire, UK
-------------------------------------------------------------------------------
What's the point of having opinions if they get ascribed to someone else?
-------------------------------------------------------------------------------roskos@csed-1.UUCP (01/15/88)
In article <11238@brl-adm.ARPA>, mcvax!kernel!mjh@uunet.UU.NET (Mark J. Hewitt) writes: > As far as I know, there is no UUCP protocol manual because the protocols > are proprietary to AT&T. I have always worried that AT&T considers the protocols proprietary, and thus won't tell people details of them, but the protocols *are* published. The 'g' packet protocol is published in a paper titled "Packet Driver Protocol" by G. L. Chesson; the paper was supposedly handed out at a Usenix presentation by Chesson several years ago, and an nroff source for it was published on the mailing list about UUCP which is sent out by tanner@ki4pv in Florida. The higher-level protocol (the one with the 'S', 'R', etc. records) was published on the Usenet about 2 years ago by Chuck Wegryzyn. However, the Chesson paper is not complete in that it doesn't give details of how error handling, etc. works, it just gives general allusions. (You will find this out if you implement your own 'g' driver!)
dennisb@tsdiag.UUCP (01/19/88)
> The 'g' packet protocol is published in a paper titled "Packet Driver > Protocol" by G. L. Chesson.... , and an nroff source for it was published > on the mailing list about UUCP which is sent out by tanner@ki4pv in Florida. I'm relatively new to the net. How doing I go about getting a copy of this from the UUCP mailing list?????? Sincerely, Dennis -- Dennis Lee Black Concurrent Computer Corp Cust Serv, Data Comm Sup 40 Pine Street Neptune, New Jersey 07756
rhes@igloo.UUCP (Richard H. E. Smith II) (01/19/88)
In article <11238@brl-adm.ARPA> mjh@uunet.UU.NET writes: >As far as I know, there is no UUCP protocol manual because the protocols >are proprietary to AT&T. I hope not, since that puts other, net-distributed programs like UUPC in trouble. The original questioner (pointer now lost, I guess) might look at UUPC if he doesn't have a source license (or doesn't want to break any rules). ---------------------------------- Dick Smith ...ihnp4!ddsw1!igloo!rhes
brian@sdcsvax.UCSD.EDU (Brian Kantor) (01/21/88)
About a year ago, John Gilmore posted a collection of uucp protocol information to the net. I've just reposted that collection to the comp.doc newsgroup. People interested in the uucp protocol will find much useful information there, although it's a collection of stuff and not a single document. If you don't get comp.doc, you should. Occasionally it's useful. - Brian Kantor (moderator of comp.doc)
chet@mandrill.CWRU.Edu (Chet Ramey) (01/22/88)
In article <364@igloo.UUCP> rhes@igloo.UUCP (Richard H. E. Smith II) writes: >In article <11238@brl-adm.ARPA> mjh@uunet.UU.NET writes: >>As far as I know, there is no UUCP protocol manual because the protocols >>are proprietary to AT&T. >I hope not, since that puts other, net-distributed programs like UUPC in >trouble. It is my impression that the source code that implements the protocols is AT&T's trade secret, not the protocols themselves. John Gilmore's packet of information was posted recently to comp.doc; it has been posted before, with no complaint from AT&T. When John posted his first attempt at a public-domain UUCP (UUSLAVE), one of the documents included was a letter from an AT&T employee stating that the source code, which used the protocols in question, was not derived from proprietary AT&T code and was therefore in the public domain. From all this I conclude that the protocols are in the public domain. I'd guess that no other "protocol manual" other than Greg Ghesson's paper has ever been written because of the misconception about the status of the protocols. Hope this is correct, Chet Ramey
nevin1@ihlpf.UUCP (01/23/88)
Recently (I think 1/20/88), a UUCP protocol information potpourri was posted to comp.doc. This may contain the information you need. (I have not had time to look at it yet.) -- _ __ NEVIN J. LIBER ..!ihnp4!ihlpf!nevin1 (312) 510-6194 ' ) ) "The secret compartment of my ring I fill / / _ , __o ____ with an Underdog super-energy pill." / (_</_\/ <__/ / <_ These are solely MY opinions, not AT&T's, blah blah blah