From: Digestifier To: Linux-Development@senator-bedfellow.mit.edu Reply-To: Linux-Development@senator-bedfellow.mit.edu Date: Wed, 21 Sep 94 16:13:21 EDT Subject: Linux-Development Digest #201 Linux-Development Digest #201, Volume #2 Wed, 21 Sep 94 16:13:21 EDT Contents: Re: "guide" (xviews) for Linux ? (Hal N. Brooks) Re: Shared Libs: working toward a permanent solution? (Alan Cox) Re: NIT for Linux (Alan Cox) Re: Shared Libs: working toward a permanent solution? (Joachim Schrod) Re: Network driver for BOCA VLB Ethernet adapter (Donald Becker) Multiple mode problem (Dan Saunders) svgalib broken in 1.1.50/51? (Bob) Kernel Goals? (David Jeske) Re: Alpha Linux (Ralf Baechle) Re: 680x0: separate newsgroup? (Michael Neuffer) Re: X.25 support ....exist ? (Rob Janssen) iBCS Great Job!, kernel version, (was Re: Where is iBCS docs?) (Keith Smith) X.25 support ....exist ? (paolo bertona) Re: Proteon Token-Ring Driver Survey (Arindam Banerji) Re: Why was ncp removed in 1.1.48? (jonathan allen) Re: Porting applications to TERM (Olaf Titz) gate syslogd: recvfrom unix: Bad file number,1.1.50 crash (Mark Swanson) Drivers for Zoom VFP14.4V (Mike Welch) SIGFPE with atof() (David Barr) Re: ADA (Grant Edwards) PCI ATM adapters for Linux ? (Werner Almesberger) AHA1542A and Linux-1.1.50+ are having troubles.. (Matti Aarnio) Re: Digi Intelligent Boards? (Troy DeJongh) EDI and linux (Electronic Data Interchange) (Tim Bass (Network Systems Engineer)) ---------------------------------------------------------------------------- From: hal@pollux.cs.uga.edu (Hal N. Brooks) Crossposted-To: comp.os.linux.help Subject: Re: "guide" (xviews) for Linux ? Date: 21 Sep 1994 09:20:58 GMT Reply-To: hal@pollux.cs.uga.edu (Hal N. Brooks) In article eedksc@aachen.ericsson.se writes: >>>>>> "Andrew" == Andrew Davison writes: > Andrew> Do anyone know of a linux port of the "guide" GUI builder > Andrew> for XView ? I know that Sun versions exist, and am a > Andrew> little concerned that it might be a Sun commercial > Andrew> venture. >The 'Guide' UI-Builder is a commercial product from Sun, I doubt that >there is a Linux version at all. Agreed. But I wanted to point out that it's OK to distribute/compile/use Sun's small libguide.a necessary to link *some* of the DevGuide generated code. I know that someone announced placing a binary of the library on sunsite some time ago, but I've never tried it. Also, if you have access to DevGuide on a Sun, then you should have access to the source for libguide.a, too. The bottom line is that although you can't run the "guide" client under Linux, you can use the code that it generates to create an app that will run under Linux. Actually you can do alot with "guide" without ever needing to link with "libguide.a" ... drag and drop is one thing that I know of that needs the library. Follow-ups to comp.os.linux.help, since this isn't a kernel related development issue. -hal ------------------------------ From: iialan@iifeak.swan.ac.uk (Alan Cox) Subject: Re: Shared Libs: working toward a permanent solution? Date: Wed, 21 Sep 1994 10:11:25 GMT In article <35dpmp$qih@news.cais.com> ericy@cais.cais.com (Eric Youngdale) writes: > People keep forgetting that to achieve the current shared >libraries that we had to add a "filter" which rewrites some of the >assembly code and adds extra indirections to any access to global >variables. Before we can use any register, we need to save the old one on >the stack, and then it needs to be restored when we are through. Also, >rewriting assembly in this way this is hard to optimize for performance, >so I would have a hard time saying which way would be faster and which >way would be slower. Thus the only consideration I have is how >actual applications perform using ELF and a.out. The timings I quoted were fixed versus elf libraries on a SYS5.3 end target. There are real differences between PIC code and magic for global variables. The libraries use few globals and things like SVGAlib already load these into stacked variables early on because of the otherwise present performance hit. If you can manage enough magic to get a 0% slowdown then you are one up on most commercial vendors I've checked against (but that won't be the first time the EYC has done that 8)) and I'm happy to accept it. There are still a small number of programs I build static that I expect I will continue to do this way - stuff like Crack that are already worst case tests of the worst cases of the C library. Alan -- ..-----------,,----------------------------,,----------------------------,, // Alan Cox // iialan@www.linux.org.uk // GW4PTS@GB7SWN.#45.GBR.EU // ``----------'`----------------------------'`----------------------------'' ------------------------------ From: iialan@iifeak.swan.ac.uk (Alan Cox) Subject: Re: NIT for Linux Date: Wed, 21 Sep 1994 10:21:54 GMT In article onders@netcom.com (Timothy E. Onders) writes: >Is anyone out there working on the Berkely Network Interface Tap for >Linux? If so, I'm interested in helping. If not, I'm interested in >attempting it. Bit pointless - there is sock_packet and a usermode implementation of BPF (see the tcpdump code). Alan -- ..-----------,,----------------------------,,----------------------------,, // Alan Cox // iialan@www.linux.org.uk // GW4PTS@GB7SWN.#45.GBR.EU // ``----------'`----------------------------'`----------------------------'' ------------------------------ From: schrod@iti.informatik.th-darmstadt.de (Joachim Schrod) Subject: Re: Shared Libs: working toward a permanent solution? Date: 21 Sep 1994 13:59:32 GMT In article <35ng6j$80l@renux.frmug.fr.net>, rene@renux.frmug.fr.net (Rene COUGNENC) writes: > Ce brave Albert D. Cahalan ecrit: > > > Every bit counts. 3% here, 2% there, 7% somewhere else - it adds up. > > Actually, it multiplies up, which is worse. > > And don't forget that there are many slow 386 boxes running Linux... I don't forget that I don't add bits but kilo- or megabytes since I don't use shared libs. Why? It's a hassle to create and maintain them. Joachim -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Joachim Schrod Email: schrod@iti.informatik.th-darmstadt.de Computer Science Department Technical University of Darmstadt, Germany ------------------------------ From: becker@cesdis.gsfc.nasa.gov (Donald Becker) Subject: Re: Network driver for BOCA VLB Ethernet adapter Date: 19 Sep 1994 18:28:18 -0400 In article <35aa16$2p7@raffles.technet.sg>, Joseph Doo wrote: >Is there a driver for this adapter? You shouldn't be asking questions like this in c.o.l.d. Or, if you do, you should asking them in the form "I was thinking about writing a driver for xxx..." That said, the updated LANCE driver in 1.1.50 has support for the PCnet/VLB chip used on the Boca BEN-VL board. I also have several minor fixes for the driver that I hope to have ready for 1.1.52. -- Donald Becker becker@cesdis.gsfc.nasa.gov USRA-CESDIS, Center of Excellence in Space Data and Information Sciences. Code 930.5, Goddard Space Flight Center, Greenbelt, MD. 20771 301-286-0882 http://cesdis.gsfc.nasa.gov/pub/people/becker/whoiam.html ------------------------------ From: saund921@cs.uidaho.edu (Dan Saunders) Subject: Multiple mode problem Date: 19 Sep 1994 22:32:07 GMT Starting at 1.1.45 (I think) my system would hang when accessing the second hard drive. I have a seagate the does support multiple block (my first hard drive), and a quantum that does not. When I compiled the kernel with the default multiple block of 64, and access my second hard drive, the controller would continually reset. This would cause me to reset my machine. Control-alt-del would not work, the system was just continually accessing the drives but it didn't seem to be doing anything. I changed hd.c to force the controller, when reset, to set the first drive at 64 and the second at 0. I could not find a way to do this except to change part of the code. Am I missing something about the way to activate the multiple block, or am I just stuck with my patch. Thanks for the help. -- =========================================================================== * Dan Saunders | What the heck I'll just leave this blank * * University of | * * Idaho, Moscow | * * saund921@uidaho.edu | * =========================================================================== ------------------------------ From: bob@xnet.com (Bob) Subject: svgalib broken in 1.1.50/51? Date: 21 Sep 1994 00:03:46 GMT i just noticed that svgalib does not seem to be working under 1.1.50 and 1.1.51. more specifically, sasteroids complains that it can't open /dev/mem im not familiar enuf with this part of the kernel to even think about figuring it out. im running Gateway P5-66 with aha1542 and linux 1.1.51 and also generic 486DX266 with ultrastor 34F and linux 1.1.50. neither one works. any ideas? bob -- Bob Hollinger <-------------------------------------> 4B Chaucer Ln. > \ bob@interaccess.com / < Disclaimer: Streamwood, IL 60107 > \/\/ bob@xnet.com \/\/ < 708-483-9391 <------------------------------------->I don't know her! ------------------------------ From: jeske@ux4.cso.uiuc.edu (David Jeske) Subject: Kernel Goals? Date: 18 Sep 1994 03:49:36 GMT A few months ago the SLS author did something that (IMHO) was rather noble and very needed. He broke the kernel up into a completely modular system where all the drivers were external and loaded. Yes, it was not the most elegant system, and yes, it still had some design inconsistancies with regard to sound drivers and the DMA memory. Yes, it still did not do a great deal to make sure that drivers from different kernels would work togeather. However, it was a start. There were many reasons that the "kernel gurus" did not integrate this into the Linux kernel proper. Some of these I know.. many of these I dont. What I want to know is: Is there some plan, or even hope, for the kernel to get cut up at least enough to allow drivers to be separated from being "kernel version dependent" at least as much as possible? Does someone working on the Linux kernel "progress" want to eventually put togeather a system of loadable drivers? I heard some rumors, that there were plans to redo the work that the SLS author did and integrate it into the Linux kernel in a "better" way. Is this happening? ver? Or, is the best way for me to see this happen to follow the Linux VIPER project? Why am I interested in this? Because I envision a day where Linux would consist of a driver set which are stable, with occasional changes, and a small "kernel" which loads things, not necessarily in a microkernel fashion, but at least enough to allow there to be a "Linux driver" and not a "Linux 1.1.47 driver". Maybe this is in the plan and all this multiple-platform stuff is just taking priority. I'm sure someone is going to tell me to go ask this on the mailing list, and if I do not get sufficient information from this post, then I will. -- David Jeske(N9LCA)/CompEng Student at Univ of Ill at Cham-Urbana/NeXT Programmer CoCreator of the GTalk Chat Software System - online at (708)998-0008 jeske@uiuc.edu (NeXTMail accepted) ------------------------------ From: ralf@resi.waldorf-gmbh.de (Ralf Baechle) Subject: Re: Alpha Linux Date: 21 Sep 1994 14:54:14 GMT In article <35npoa$g3i@nntp.Stanford.EDU>, dhinds@allegro.stanford.edu (David Hinds) writes: |> Kevin Marcus (datadec@corsa.ucr.edu) wrote: |> |> : Actually, the Alpha suffers significant performance penalties when dealing |> : with 32 bit vs. 64 bit quantities. (Yet another reason NT is slow on |> : the lower end Alpha's). |> |> Are you sure about this? In my experience, on 64 bit architectures |> (like MIPS R4x00, IBM POWER, Alpha, etc) using 32 bit quantities is |> often significantly faster than going with the full 64 bits, probably |> due to better cache utilization. You're right at least for MIPS R4x00. On 64 bit MIPS CPUs there is no difference in execution time between 32 or 64 bit code for all integer instructions (exceptions are multiply and divide instructions). However using 64 bit memory operations where 32bit would suffice may result in a performance penalty due to the cache utilisation. Even worse, some existing board designs only have a 32 bit bus from memory to the second level cache. That's just one of the reasons why Linux/MIPS is 32 bit for now. Full 64 will follow, of course. Ralf ------------------------------ From: neuffer@wilbur.zdv.Uni-Mainz.DE (Michael Neuffer) Subject: Re: 680x0: separate newsgroup? Date: 19 Sep 1994 17:06:16 GMT Kielhorn (c0031098@ws.rz.tu-bs.de) wrote: : There is a german group : maus.os.linux68k : (Language is still german but may/will change if there are more : English speakers) : Main subject: Linux on Atari TT and Falcon That's only because are pretty quiet right now. Ralf for example is working on the MIPS port at the moment. He still answers questions. People interested in the mips port are also welcome there. Mike -- Maus-/UseNet:Michael_Neuffer@wi2.maus.de Usenet :neuffer@goofy.zdv.uni-mainz.de Fido :Michael Neuffer@2:245/5530.5 ------------------------------ From: rob@pe1chl.ampr.org (Rob Janssen) Subject: Re: X.25 support ....exist ? Reply-To: pe1chl@rabo.nl Date: Wed, 21 Sep 1994 08:35:08 GMT In evansmp@mb4715.aston.ac.uk (Mark Evans) writes: >paolo bertona (bertop@c700-1.sm.dsi.unimi.it) wrote: >: I am searching for some support for X.25 cards >: under Linux, can somebody help me ? >The only X25 support is the AX25 written by Alan Cox, >You might like to look at this code for ideas, I >suspect you will find that things like the HDLC >management you can leave to the hardware. Be warned that "AX.25" has no other relation to "X.25" than that they both use an LAPB link layer. The name "AX.25" was probably chosen at a time when the designers still thought that the X.25 model would be used. They first implemented the link level, and the X.25 PLP was left for later. But is was never really implemented for use by end-users (only one network product uses it internally). So, AX.25 code does not buy you much when you want to have X.25... Rob -- ========================================================================= | Rob Janssen | AMPRnet: rob@pe1chl.ampr.org | | e-mail: pe1chl@rabo.nl | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU | ========================================================================= ------------------------------ From: keith@ksmith.com (Keith Smith) Subject: iBCS Great Job!, kernel version, (was Re: Where is iBCS docs?) Date: Wed, 21 Sep 94 11:53:01 GMT In article , Al Longyear wrote: >Instructions for installing and using iBCS is in the package on tsx-11. Well mostly anyway. It seems that every time I install something for linux I have to make a leap of faith and draw on experience. :) I guess that's why it's fun? :) It would be _nice_ if developers would put the linux revision the code was compiled under. I grabbed the iBCS code from September and basically added patches to a kernel source tree re-compiling in-between until I _thought_ I had it with 1.1.39. Except that didn't fly either, on up to 1.1.42 (mentioned in passing in the ChangeLog) and viola, insta-SCO box. Simply Amazing. Flawless execution of some KEY SCO programs. I have another commercial program that runs under Linux/iBCS not listed: BBx/4. Great job, I owe you a beer. -- Keith Smith Digital Designs keith@ksmith.com 5719 Archer Rd. PO Box 85 Free Usenet News and Internet Mail Services Hope Mills, NC 28348-0085 All 28K/14K Modems (910) 423-4216/7389/7391 Somewhere in the Styx of North Carolina ... 14K-V.32/28K-V.34/28K-V.34 ------------------------------ From: bertop@c700-1.sm.dsi.unimi.it (paolo bertona) Subject: X.25 support ....exist ? Date: 19 Sep 1994 19:23:46 +0200 I am searching for some support for X.25 cards under Linux, can somebody help me ? Greetings Paolo Bertona ------------------------------ From: axb@defender.dcrl.nd.edu (Arindam Banerji) Subject: Re: Proteon Token-Ring Driver Survey Date: 21 Sep 1994 15:37:23 GMT The author is Peter Schrijver ( stud11@cc4.kuleuve.ac.be) - you might want to contact him. He's been working on an MCA version - I'm sure that other people have moved it to ISA bus machines, by now. -thanx ============================================================================= Arindam Banerji (219)-631-5273 (Voice) 384 FitzPatrick Hall (219)-631-5772 (Voice) Dept. of Computer Science and Engineering (219)-273-0862 (Voice) University of Notre Dame (219)-631-9260 (FAX) Notre Dame, IN 46556 axb@cse.nd.edu (E-mail) HomePage:http://www.nd.edu/Departments/EN/CSE/DCRLab/HomePages/axb ============================================================================= ------------------------------ From: jonathan@mirror.demon.co.uk (jonathan allen) Subject: Re: Why was ncp removed in 1.1.48? Reply-To: jonathan@mirror.demon.co.uk Date: Fri, 16 Sep 1994 09:40:15 +0000 In article iialan@iifeak.swan.ac.uk "Alan Cox" writes: > It was something Mark Evans added that really didnt belong in the kernel > (NCP uses three sockets in a way you can handle cleanly in user mode). For the sake of information, and it might be useful to me, what is/was NCP, and where can it now be found ? Jonathan =============================================================================== Jonathan Allen | jonathan@miror.demon.co.uk | Voice: 0271-79023 Barum Computer Consultants | jeremiah@cix.compulink.co.uk | Fax: 0271-24183 =============================================================================== ------------------------------ From: uknf@rzstud1.rz.uni-karlsruhe.de (Olaf Titz) Subject: Re: Porting applications to TERM Date: 21 Sep 1994 13:23:53 GMT > >Does anyone have some decent documentation on what is required to port > >applications to TERM. Any help would be appreciated. > As Term-HOWTO maintainer I'm interested too. Porting software is still a major > gap in the HOWTO which needs to be filled. Porting shouldn't be necessary with the new termnet library. Add in the program Makefile following definitions: to CFLAGS -include /usr/local/lib/termnet.h (or whereever) to LIBS -ltermnet -L/usr/local/lib ( ="= ) That should do it. Note that -include is not understood by some C compilers (but gcc does grok it :-) Ideally, programs compiled in this way run with and without term and/or TCP/IP, they always take what's available. Olaf -- ___ olaf@bigred.ka.sub.org - uknf@rz.uni-karlsruhe.de __ o click __/<_ also: s_titz@ira.uka.de - uknf@dkauni2.bitnet - praetorius@irc _)>(_)_________ "now i find that most of the time love's not enough in itself" ------------------------------ From: ag010@FreeNet.Carleton.CA (Mark Swanson) Subject: gate syslogd: recvfrom unix: Bad file number,1.1.50 crash Reply-To: ag010@FreeNet.Carleton.CA (Mark Swanson) Date: Mon, 19 Sep 1994 22:37:06 GMT I get the error message in the Subject line, then I get: gate last message recieved 78732 times gate last message recieved 305048 times gate last message repeated 1549763 times (all repeated, sorry) gate last message repeated 1565857 times etc. etc. root can no longer log in. No one can log in from serial ports. Rebooting is the only answer as syslogd is taking up all of the processing power. I have only changed mgetty to 0.2. I've changed it back to 0.16. If you have any owther ideas, I'd like some help. Thanks. -- Mark Swanson. ag010@freenet.carleton.ca Linux FREE BBS V.Fast 24,000bps. Ottawa, Canada. (613)-829-1941 ------------------------------ Reply-To: mike@muppetski.win-uk.net (Mike Welch) From: mike@muppetski.win-uk.net (Mike Welch) Date: Tue, 20 Sep 1994 13:06:56 GMT Subject: Drivers for Zoom VFP14.4V Does anyone have a sound driver for the Zoom VFP14.4V, alternatively can someone tell me how I can use it`s audio features since the manual is rather scant in the programming area. Cheers Mike ------------------------------ From: barr@pop.psu.edu (David Barr) Subject: SIGFPE with atof() Date: 19 Sep 1994 13:53:06 -0400 I'm running Slackware 2.0, and i'm trying to compile a program called xweather. (You can get the program from ftp.pop.psu.edu, files /pub/src/xweather.tar.Z and /pub/src/xweather.patch1) The code runs fine on a SPARC, under both 4.1.3 and under Solaris. Friends of mine have it working on other platforms as well. If I compile it under Linux (running either 1.0.9 or 1.1.51), I get a SIGFPE at an atof(). The string atof() is reading is a valid number. I'm using gcc 2.5.8, on a 386 with a 387. The funny thing is that it doesn't seem to depend on which number it's trying to convert to a float, it will bomb after a certain number of atof()'s are called. While I'm at it, when I compile it, I end up with a static binary. I have the shared X libs loaded as well as the static ones. Why isn't a shared binary being built? --Dave ------------------------------ From: grante@reddwarf.rosemount.com (Grant Edwards) Subject: Re: ADA Date: Mon, 19 Sep 1994 21:58:27 GMT Darren Davenport (davenpor@pat.mdc.com) wrote: : There is no ADA compiler for Linux, but there is an Ada9x compiler. :-) Thank you Darren; that's one of my pet peeves. I could never figure out why anybody would want compile the American Dental Association. The language in question is Ada not ADA. It's not an acronym it's a woman's name. When I was in school a word that wasn't properly capitalized was counted as misspelled. I can't stand it when people write about PASCAL either. Let's not even talk about how the word "nuclear" is pronounced. -- Grant Edwards |Yow! Thank god!!.. It's Rosemount Inc. |HENNY YOUNGMAN!! | grante@rosemount.com | ------------------------------ From: almesber@bernina.ethz.ch (Werner Almesberger) Crossposted-To: comp.dcom.cell-relay Subject: PCI ATM adapters for Linux ? Date: 21 Sep 1994 14:06:33 GMT I'm looking for ATM adapters with the following characteristics: - 155 Mbps, fiber (OC-3) - for PC hardware, PCI bus - usable for Linux The last point probably implies that somebody (maybe I) has to write a device driver. Thus, the necessary specifications need to be available and free from non-disclosures, etc. I'd be most thankful for information on available products or imminent releases, or possibly on people already doing work in that direction. I'll post a summary if I get replies that aren't also posted to the newsgroup(s). - Werner -- _________________________________________________________________________ / Werner Almesberger, ETH Zuerich, CH almesber@bernina.ethz.ch / /________________________________________________________________________/ ------------------------------ From: mea@utu.fi (Matti Aarnio) Subject: AHA1542A and Linux-1.1.50+ are having troubles.. Date: Wed, 21 Sep 1994 18:20:16 GMT On System: - i486/33 32MB DRAM (ISA) - Adaptec AHA1542A + 600 MB CDC Wren-V disk When running Linux 1.0.9 there is no trouble, however with Linux 1.1.50 and 1.1.51 there appear spodaric data corruption on the disk WRITES. The corruption follows rule: At a 1k boundary write 8 bytes of 0xFF instead of the true data Location of that error varies within a file, but it is always on bytes N*1k..N*1k+7 The error may appear at multiple spots of the file also. At changes46 (for 1.1.46) there is a comment about slow DMAs, and unnecesserity to be slow at all times.. Might some slowdown be adviced in here ? I think this same applied to Linux-1.1.18, as I have seen similar weirdos at other aha1542a using machine... /Matti Aarnio ------------------------------ From: troyd@digibd.digibd.com (Troy DeJongh) Subject: Re: Digi Intelligent Boards? Date: Mon, 19 Sep 1994 14:07:57 GMT ahmed@oea.xs4all.nl (Ahmed Naas) writes: >madderra@myhost.subdomain.domain wrote: > ^^^^^^^^^^^^^^^^^^^^^^^ >You need to change this :-) >: I visited the Digiboard display at Interop this week, and was told by one >: of the sales reps that Digiboard was getting pressure from SCO not to >: promote the creation of new drivers for Linux. In particular, I was >: asking about thier new ISDN board they had on display. >That might be true, but what possible leverage does SCO have over Digi? >: --Bob >-- >The above is a result of random neuron activity in the writer's brain. >Ahmed M. Naas ahmed@oea.xs4all.nl >---------------------------------------------------------------------- Well, I'm responsible for the maintenance/development of our SCO serial drivers and the development of the Linux driver for our PC/Xe product, and I haven't heard anything like that yet. Plans are to release a Linux driver for our PC/Xe line of products this fall. -- Troy De Jongh Digiboard (troyd@digibd.com) ------------------------------ From: bass@cais2.cais.com (Tim Bass (Network Systems Engineer)) Subject: EDI and linux (Electronic Data Interchange) Date: 21 Sep 1994 01:46:49 GMT Seems like EDI (Electronic Data Interchange) for buying and selling, especially with the US government is really heating up. Is anyone doing anything with linux for EDI?? ------------------------------ ** 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 ******************************