From: Digestifier To: Linux-Development@senator-bedfellow.mit.edu Reply-To: Linux-Development@senator-bedfellow.mit.edu Date: Tue, 21 Sep 93 03:13:17 EDT Subject: Linux-Development Digest #114 Linux-Development Digest #114, Volume #1 Tue, 21 Sep 93 03:13:17 EDT Contents: Re: Linux, Notebooks, XFree86, and LCDs (James S. Vera) Re: Linux ISDN code under development? (Matthias Urlichs) Re: Rlogin, telnet, etc... (SLS Linux 1.03/NET-2) Signal problems? (Dan Wilder) Re: No smart serial boards??? (Jon Brawn) [Q] Problems porting XF2.2 to Linux (Weimin Zhao) [Q] SCSI tape (auto sense between QIC-24 and QIC-11) (W. Tait Cyrus) Tandberg 3600 and its handshaking problems (was Re: SCSI Timeouts) (Ed H. Chi) Re: To all device driver writers; boot-time messages. (Michael A. Irons) Re: EXT2 fs bug? (Michael A. Irons) term (Kenneth H. Simpson) Re: What do people think about /config? (Erik Bos) ---------------------------------------------------------------------------- Crossposted-To: comp.os.linux.help,comp.windows.x.i386unix From: vera@fanaraaken.Stanford.EDU (James S. Vera) Subject: Re: Linux, Notebooks, XFree86, and LCDs Date: 21 Sep 93 00:50:23 GMT pcolsen@super.org (Peter C Olsen) writes: >I'm looking for anyone who has ever successfully installed Linux and >XFree86 on an LCD laptop. I have a laptop that I carry for school and >Reserves and I have been told that getting XFree to run on an LCD will >be non-trivial. >I have read the FAQ. I'm using XFree86 version 1.3. The mono server >runs on my laptop driving an external monitor but does not drive the >LCD. I have tried Xega, but cannot get it to run. I'm willing to >invest substantial time, but I would like an existance proof before I >do so. I run the mono server of XFree version 1.2 on my Compaq LTE386s/20 notebook. It installed with no problems. I tried Xega but its font support seemed kinda clunky and I can live with mono for now. I can't switch to an external monitor as that is done through software and I haven't gotten around to finding the necessary BIOS call... But, XFree on an LCD does exist and, for me, was trivial. Your milage may vary. -- James S. Vera | Internet | Voice: +1.415.723.1089 Stanford University | vera@anna.stanford.edu | FAX: +1.415.725.7398 Are these windows going to change my original document?-Time/Life dweeb ------------------------------ From: urlichs@smurf.sub.org (Matthias Urlichs) Subject: Re: Linux ISDN code under development? Date: 21 Sep 1993 00:52:06 +0200 In comp.os.linux.development, article <271c91$lqf@inasys.inasys.uucp>, behe@inasys.uucp (Bernd Hentig) writes: > In article <26ero6$1hu@smurf.sub.org> urlichs@smurf.sub.org (Matthias Urlichs) writes: > >In comp.os.linux.development, article <25r3nb$dtc@panix.com>, > > rwebb@panix.com (Russell Webb) writes: > >> Is any work currently being done on code for Linux to support a > >> ISDN adaptor? I saw some talk of it months ago, but all such > >Yes. > ... > >The current incarnation is also Streams-based -- this is the real thing, > >which means that somebody will have to sit down and rewrite all the sticky > >AT&T code before it can be released. I fear that this somebody will be me, > >though I'd _really_ appreciate if somebody else could take that load > >off me. > ??? Free or private code ??? > The final code will probably be copylefted. > I didn't see any ISDN drivers on the major Linux archive sites yet - > is it still pre - alpha ??? As I said, it contains AT&T code and thus CANNOT be posted to any major archive sites! > Which boards are supported ? I understand most ISDN boards have > a proprietary low-level interface (i.e. manufacturers won't tell > you about registers and commands ...) ... but at least with boards that don't have their own CPU, such things are easily figoureoutable. > >The other problem is that the Linux TTY code isn't very modular, and > >therefore it'll be well nigh impossible to log in through ISDN [...] > > As far as I understand, e.g. the telnet protocol is caring about things > like this ... I wonder if it is that important to support tty emulation > via ISDN if you could have full networking support (SLIP) ... SLIP currently needs the tty interface, thus the current Linux SLIP won't run directly on ISDN unless the tty interface is reorganized. Routing IP packets through the PTY interface is horrible. In fact, I'd like to avoid going through a PTY for straight ISDN logins too if at all possible. Not everybody can or wants to do TCP/IP over their ISDN link. -- Matthias Urlichs -- urlichs@smurf.sub.org -- Phone: NONE; use email or lose. Schleiermacherstrasse 12 -- 90491 Nuernberg -- Germany || Linux+Mac Consulting ------------------------------ From: danw@hebron.connected.com (Dan Wilder) Subject: Re: Rlogin, telnet, etc... (SLS Linux 1.03/NET-2) Signal problems? Date: 17 Sep 1993 09:34:56 -0700 spj@ukelele.gcr.com (Guru Aleph_Null) writes: >Does anyone know if the rlogin and telnet that come with SLS Linux >1.03 have signal handling problems (e.g. I hit ^C to kill a process >remotely, instead, my rlogin/telnet is killed!)? Its driving me batty! >This is with bash 1.12.1 that is on my system. >I suspect it's the telnet/rlogin binaries (and probably with ftp, but >I don't remember testing for this behavior with FTP.) I'm using them >over a 14.4 SLIP link. So telnet does this too, now? I'm running 0.99pl9-2 and it drives me batty with rlogin, but telnet does OK. This is over ethernet to SCO unix. The ^C problem has been mentioned in other posts, I had hoped it would have been fixed by current release. Guess not. --- Dan Wilder ------------------------------ From: jonb@specialix.com (Jon Brawn) Subject: Re: No smart serial boards??? Date: Tue, 21 Sep 1993 03:08:15 GMT bogart@ucsee.Berkeley.EDU (Ken Geis) writes: > I read in the (I believe) serial FAQ that no drivers exist for >intelligent serial multi-port boards. The FAQ also mentioned that they >would likely ever exist. Can anyone explain to me why? I'd like to >run a good size BBS from a Linux box, but I won't start writing it until >I know it can work (I have to get some more programming experience too, >but that's beside the point :) ). Thanks in advance, Putting my pompous^H^H^H^H^H^H^Hprofessional head on again, let me explain: Hardware: Physically speaking, at the level of electrons and things, there is no reason why intelligent serial I/O cards can't be installed in a machine running Linux. Software: There is no reason why a driver for an intelligent serial I/O card couldn't be written: drivers have been ported to such hideously different environments as SCO Unix, Microsoft Windows NT, IBM AIX 3, MS-DOS, ... so it is by no means impossible to imagine this happening. Cost: Financially, we start getting into an area whereby people get a little reluctant to pay $1500 for a 16 port I/O solution when most of their kit was free. Example prices of Specialix I/O solutions: Dumb: I/O4 AST four port look-alike $ 300 Semi-intelligent: I/O8+ Semi-intelligent 8 port card $ 495 Intelligent (Z280 CPU + dumb octal UARTs) SI/4 Intelligent 4 port solution, host card + 4 port module $ 795 SI/8 Intelligent 8 port solution, host card + 8 port module $ 995 SI/16 Intelligent 16 port solution, host card + 2 x 8 port modules $ 1,495 SI/32 Intelligent 32 port solution, host card + 4 x 8 port modules $ 2,495 Intelligent (Z280 CPU + intelligent quad UARTs) XIO 8 Intelligent 8 port solution, host card + 8 port module $ 1,145 XIO 16 Intelligent 16 port solution, host card + 2 x 8 port module $ 1,795 XIO 32 Intelligent 32 port solution, host card + 4 x 8 port module $ 3,095 Intelligent, distributed modules (T225 Transputers, intelligent UARTS) RIO 32 Distributed 32 port solution $ 3,995 RIO 64 Distributed 64 port solution $ 7,195 RIO 128 Distributed 128 port solution $13,595 RIO 512 Distributed 512 port solution $54,380 The other intelligent serial I/O card vendors all have products with roughly the equivalent performance/price ratios, some are better, some not as good, but most are in the same ballpark figure (suprise) Legal: Legally, there are real problems. The GPL's restrictions are so severe that most companies are *unable* to release drivers for their products, simply because of the uncertainty of the legal position wrt the GPL. [I think I mean GPL - that rather savage horned and toothed software license that Linux is covered by]. If the requirement is that software source is made available, then there is *no way* that most of the I/O vendors are willing to do that. I have been toying with the idea of writing drivers for the I/O8+, SI range and XIO range, however, I am always put off by the software license. We aren't about to publish our source code. Support: When you call us with a support call, it costs us real money to answer the call. Each person who calls us up prevents one member of our support team from (i) helping someone else (ii) writing up a report of problems already seen, (iii) finding workarounds and (iv) going to lunch. Politics: It's very hard to justify writing device drivers for an operating system that is either unknown or regarded as being a hackers toy by management. I am perfectly willing to write device drivers for Linux for Specialix I/O cards. Unfortunately, they would probably never get out of our door. Can anyone answer the following question for me: I would like to distribute just a binary of the driver with a small C file to get it included into the kernel. Is this possible within the confines of the Linux software license restrictions? [ IF you think you are going to answer this question, you first of all need to convince me you know what you are talking about - I don't want opinions from 18 year old freshmen, I want facts from people who are actively involved in the FSF projects]. If you really want to hang many ports off of a Linux box, I offer you the Specialix MTS (Modular Terminal Server). The Modular Terminal Server is a high performance, flexible and secure method for businesses to expand TCP/IP networks. MTS allows terminals, printers, and modems easy access to resources of a TCP/IP network. The modular design of MTS enables from 8 to 32 users to be linked to the network. Access to the TCP/IP ethernet network is made via either thin or thick ethernet. 8 Port RS232 MTS $1695 8 Port Adaptor $ 650 So, for a 32 port terminal server, $1695 + 3 x $650 = $3645 As you can see, more expensive than direct connect serial I/O, and also much less efficient (= greater system overhead) because of having to invoke the mighty TCP/IP stack software for every character typed by the user! > Ken Jon. ------------------------------ From: wzhao@mcs.kent.edu (Weimin Zhao) Crossposted-To: comp.lang.tcl Subject: [Q] Problems porting XF2.2 to Linux Date: 21 Sep 1993 03:38:14 GMT Hi, All, I have Tcl 6.7 and Tk 3.2 installed on a Linux box. Now I'm trying to install XF2.2. After modifying Makefile and starting "make install", I got the following error messages: invalid command name: "auto_mkindex" while executing "error "invalid command name: \"$cmdName\""" (procedure "unknown" line 39) invoked from within "auto_mkindex ../autoProcedures *.t" (file "./lib/makeAutoTmplt.tcl" line 35) invoked from within "source ./lib/makeAutoTmplt.tcl" make[1]: *** [makeAutoTemplates] Error 1 (ignored) My guess is that the command auto_mkindex is somehow missing from the Tcl/Tk on the Linux. Can someone enlighten me on this? Any help/pointer is greatly appreciated. -Weimin wzhao@mcs.kent.edu ------------------------------ From: cyrus@jemez.eece.unm.edu (W. Tait Cyrus) Subject: [Q] SCSI tape (auto sense between QIC-24 and QIC-11) Date: 21 Sep 1993 04:16:32 GMT I have hooked an old tape drive up to my linux box (0.99.10) and find that I can't read any of the tapes created while the drive was attached to a Sun. I have hooked the very same drive up to a Sun and all is fine. It appears that SunOS auto senses and switches between QIC-24 and QIC-11. Anything like this for Linux? Thanks in advance. -- W. Tait Cyrus e-mail: cyrus@su.com Solutions Unlimited Phone: 719-260-7227 4710 Nightingale Dr. #M202 Colorado Springs, CO 80918 ------------------------------ Crossposted-To: comp.os.linux.misc From: ehhchi@maroon.tc.umn.edu (Ed H. Chi) Subject: Tandberg 3600 and its handshaking problems (was Re: SCSI Timeouts) Date: Tue, 21 Sep 1993 04:51:06 GMT In article apeterso@badlands.NoDak.edu (Alan Peterson) writes: >Bill Mitchell (mitchell@mdd.comm.mot.com) wrote: >: in comp.os.linux.misc, c@royle.org (Chris Royle) said: > >: >I have had a 280 MB SCSI II Seagate Hard Disc Drive running off my >: >Adaptek 1540B SCSI controller under Linux for ages with no problems at >: >all so far. >: > >: >I have now added a Sankyo CP150SE Tape streamer to the chain, and this >: >seems to be causing timeouts during my backing up with tar. (That is, >: > >: >[problems description deleted] >: >Anyone got any ideas ? >: > > >: I've been having similar problems since I added a Tandberg SCSI tape to >: my system, have posted about them previously, and just posted about them >: again in c.o.l.development in reaction to a related posting I saw there. > >I, too, am seeing timeouts with two SCSI disks and a Tandberg 3600 tape >drive. All are internal with the Tandberg at the end of the chain and >terminated. Unfortunately, I'm also seeing 'in2000_abort' with the alpha >version of the IN-2000 driver and the partition table on the second >330 meg drive (where linux lives) is trashed. I can't even do an fdisk write >without hanging the system. > >Alan Peterson >apeterso@plains.nodak.edu I have been researching into the problems with Tandberg 3600 tape drives. Apparently, they were originally designed to IBM OEM. They are not SCSI-1 or SCSI-2 devices. They are SCSI, but they are not true to the SCSI-1 or the SCSI-2 specification. Numerous people has been having problems with Tandberg 3600 drives, because they don't handshake correctly. I contacted Vidar Mortensen of Tandberg Data in Norway: Kjeisasvelen 172 P.O. BOX 133 Kjelsas N-0411 OSLO, Norway fax + 47 22 95 13 55 phone 47 22 18 90 90 email: vimo@tdata.no He informed me of the following: > I would also like to get a new firmware for my Tape drive. No offense, > but the U06 and the U07 revisions are really bad. They are barely usable > by anything. (Perhaps because of its deficiencies, it will only work with > IBM original setups.) (The 3600's were originally packaged to ship with > IBM machines.) > > PLEASE PLEASE find a new firmware for us, because we badly need it. > > Thanks. > > PS: to David, I suppose this is a real test for Tandberg's product > support. :) :) > Your Tandberg 3600 firmware (U06 or U07) are IBM OEM versions. Tandberg made this firmware after IBM spesifications. You can upgrade your firmware EPROM to either standard SCSI-1 or SCSI-2. The version numbers are: Standard SCSI-1 -07:06 Standard SCSI-2 =08:15 (Supports Windows NT) The price from Norway is NOK 350 (approx. US $ 49) + NOK 150 for freight and handling. You can send your order to me. ===========================cut =====cut =======cut================== Well, I think a whopping 70 dollars for an EPROM upgrade was way too rich for my blood. I asked if I could get it within the US, he said: Alternatively, you can connect our US Branch, and ask their prices. The address is: Tandberg Data, INC. 2649 Townsgate Rd. Suite 600 Westlake Village, CA 91361 voice (805) 495-8384 fax (805) 495-4186 I don't know how much it is from the US Branch for such an upgrade, but I would most certainly like to know if anyone finds out. Best Regards, Ed -- o/ \ / \ / / \o email: ehhchi@staff.tc.umn.edu /# ##o # o## #\ / \ / \ /o\ / |\ / \ ------------------------------ From: mirons@icarus.ci.net (Michael A. Irons) Subject: Re: To all device driver writers; boot-time messages. Date: Mon, 20 Sep 1993 04:09:26 GMT In article <1993Sep17.184413.6604@super.org>, Donald J. Becker wrote: >In article <27c5fp$17g@bambi.zdv.uni-mainz.de>, >Dominik Kubla wrote: Stuff I didn't want to comment on deleted >> This way you can >> easily upgrade drivers, create special kernels with/without certain drivers >> and commercial software/hardware developers can provide binary only drivers. > >This would be a Very Bad Thing. Luckily the GPL should protect the kernel >itself from object-only distributions. [[ Slightly off topic here...]] But if >we had a standard, published interface definition for loadable device drivers >I suspect it would take only weeks before we had proprietary, binary-only >drivers. People with a short-term perspective might not find that dangerous, >or even undesirable, but I certainly do. Shame on you Don, that is no reason for having bad or confusing code. A standard, published interface definition should be a MUST for ALL the code. Linux includes the code for the hacker, and hobbiest. The petty, self-centered people who might put out object-only distributions be damned. It would be nice to stop them, but I seriously think that easy to write for, crystal clear, well documented code will do _much_ more to help the Linux/GNU/FreeWare/ShareWare cause then stopping a few petty people. The people who really care about Linux and what it stands for won't 'buy' object-only things. Look at how the Diamond products are shuned now, even though they could be supported, because of Diamonds proprietary attitudes. I think that The learning curve for Linux code is getting outrageous. People new to Linux have a hard time following the code, writing and 'playing' with it, and that hurts the cause. Your statement sounds almost like the old programmers job-security adage "keep it confusing so they won't fire me". Personally I see this is proprietaryism in another guise. It's end result is to just shut someone out of a 'market'. I use this code, as do others, and I can't follow alot of it because there is no standard, published interface definition for just about anything. I'd rather see clean, easy to use code (and have to withstand a few self-centered fools), then have to wade through the confusion that the code is becoming. I liked Minix, where I could add a driver stub in an hour or two because the code was easy to read, and laid out well. I didn't have to be a Minix source-code guru to do it, with Linux you do :( > >-- > >Donald Becker becker@super.org >IDA Supercomputing Research Center >17100 Science Drive, Bowie MD 20715 301-805-7482 -- Mike Irons mirons@Icarus.CI.NET ------------------------------ From: mirons@icarus.ci.net (Michael A. Irons) Subject: Re: EXT2 fs bug? Date: Mon, 20 Sep 1993 20:23:59 GMT Yes - I had this problem. The count and bitmap don't agree. Get the NEWEST version of e2fsck, as previous version didn't check the last group correctly. This will fix the problem. -- Mike Irons mirons@Icarus.CI.NET ------------------------------ From: ken@kronos.arc.nasa.gov (Kenneth H. Simpson) Subject: term Date: Tue, 21 Sep 1993 05:18:09 GMT Where can I find the source for the program term ? -- ============================================================================== Kenneth Simpson NASA Internet: ken@ptolemy.arc.nasa.gov Ames Research Center, MS/269-1 UUCP: ames!ptolemy!ken Moffett Field, CA 94035-1000 ------------------------------ From: erik@trashcan.hacktic.nl (Erik Bos) Subject: Re: What do people think about /config? Date: Mon, 20 Sep 1993 22:32:00 +0911 aso@ic.sunysb.edu (Angelito So) writes: > Personally, I like the idea of having /conf and then linking the >files from there to their normal *nix area. This made my life a whole lot >easier when I re-installed the slackware distribution and just restored the >links since I have everything setup in /conf like this: > /conf > ----/conf/X11 > ---------/conf/X11/app-defaults > ----/conf/net > ... >That way you can just go in and do a > ln -sf /conf/X11/app-defaults/* /usr/X386/lib/app-defaults >to restore your defaults. This also applys to other files like Xconfig, >net-2 stuff, etc... It is very handy when reinstalling/backing up a system. Unix wizards who don't like this directory can still directly edit the files in /etc and other directories without using /conf. The Linux/PRO distribution I'm using has got directories in /conf for X11,net,mail,news,uucp. -- Erik Bos (erik@trashcan.hacktic.nl) ------------------------------ ** FOR YOUR REFERENCE ** The service address, to which questions about the list itself and requests to be added to or deleted from it should be directed, is: Internet: Linux-Development-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux.development) via: Internet: Linux-Development@NEWS-DIGESTS.MIT.EDU Linux may be obtained via one of these FTP sites: nic.funet.fi pub/OS/Linux tsx-11.mit.edu pub/linux sunsite.unc.edu pub/Linux End of Linux-Development Digest ******************************