From: Digestifier To: Linux-Development@senator-bedfellow.mit.edu Reply-To: Linux-Development@senator-bedfellow.mit.edu Date: Fri, 7 Oct 94 17:13:29 EDT Subject: Linux-Development Digest #276 Linux-Development Digest #276, Volume #2 Fri, 7 Oct 94 17:13:29 EDT Contents: Re: VESA and SVGAlib? (Christopher Wiles) Is anyone running websterd on linux? (Thaddeus H. Wood) Re: AHA-1742 driver near optimal? Ult 24F? (Mary Shenk) Re: Linux Mud (Joseph Angelis) Help With LCD, Cirrus and XFree86 (Lucas James Sheneman) Re: VESA and SVGAlib? (Phil Howard) Re: Compiling progs using port I/O (Frank Lofaro) Re: Adaptec AHA-2940 PCI SCSI card support.... (Hans de Hartog) Re: ext2fs vs. Berkeley FFS (H. Peter Anvin) Re: Telnet & ftp freeze! (Peter H. Lemieux) Re: Beautifying Linux/Xfree (DAVID L. JOHNSON) Re: Shared Libs: working toward a permanent solution? (Eric Youngdale) Xfree 3.1 and SPEA MirageP64 (Linux) (Christoph Martin) Re: Telnet & ftp freeze! (Ralph Sims) LINUX Logical volumes (Richelo Killian) Freeware Motif/Xview/tty interface library ? (Patrick Spinler) ---------------------------------------------------------------------------- From: a0017097@wsuaix.csc.wsu.edu (Christopher Wiles) Subject: Re: VESA and SVGAlib? Date: Fri, 7 Oct 1994 01:24:17 GMT hhenson@inyanga.cs.wits.ac.za (Howard P. Henson) writes: : Andy Bailey (bailey9@muvms6.wvnet.edu) wrote: : : I have what might be a dumb question, about SVGAlib and video modes. I don't : : know diddly about programming graphics drivers, but here goes. : : With most DOS applications, namely graphics viewers, instead of specifying : : the drive specific to my card, I simply use the VESA driver, and voila, all my : : cardsmodes are recognized. Would this be possible for SVGAlib? I have overheard : : bits of conversation among Linux developers about avoiding making BIOS calls ( : : I guess to ensure portability to other processors). Is the case the same here? : : The problem is not so much portability as the fact that VESA : drivers (and I asume you have a bios version) is 16 bit code and to run : it would require on to change CPU modes, and start thunking (16 -> 32 bit : conversions) etc. producing i) Slow access ii) Possible unstabilities on : the system which may land up with system crashes etc. .. things run ever so much faster when one accesses the SVGA chip in question directly ... no overhead added by going through the 8/16 bit 250ns VGA BIOS (which may contain bugs to begin with) accessed worst-case through an ISA slot at 8Mhz (I _did_ say worst-case). I don't think that the CPU has to switch modes to deal with the bus ... if that were the case, you'd switch modes every time you accessed the hard drive. -- Chris a0017097@wsuaix.csc.wsu.edu wileyc@halcyon.com wileyc@quark.chs.wa.com "... but I want to use all eight comm ports SIMULTANEOUSLY!" PGP 2.6 public key available by finger for the clinically paranoid. ------------------------------ From: thw@sentient.sentient.com (Thaddeus H. Wood) Crossposted-To: comp.os.linux.admin Subject: Is anyone running websterd on linux? Date: 7 Oct 1994 01:43:57 GMT Greets all. I'm attempting to run the websterd server on linux. I've gotten the client & server sources that David Curry wrote at Purdue. I've managed to get it compiled, but there seems to be some sort of problem using libdbm.a. It compiles and links fine, but for some reason, calls to dbm_open or dbm_fetch seem to fail. Does anyone have any sort of inkling of what I speak? Is dbm the problem here? Do I need some sort of external dbm software? I'm running a full Slackware 2.0.1 installation. Thanks. -- Thaddeus H. Wood thw@sentient.com -- -- If you're not part of the solution, you're part of the precipitate. -- ------------------------------ From: mkshenk@u.washington.edu (Mary Shenk) Subject: Re: AHA-1742 driver near optimal? Ult 24F? Date: 7 Oct 1994 01:45:38 GMT It seems that while I am hitting the disk hard on my dx2/66 w/an AHA-1742 host adapter, I miss term packets. I have a 16550 UART. Should this be happening? It seems my system slows more than I think it "should" while hitting the disk. How much work has gone into the 1742 drivers under Linux? Are they close to optimal? What about the UltraStor 24F? Will this be a superior card under Linux, due to technical superiority and/or a better driver? Drivers aside, can anyone comment on the rel. technical merits of these 2 cards? Thanks. (term is configured properly. I promise. Works wonderfully except when I hit the disk hard. These event are definitely linked, have done may trials.) ------------------------------ From: angelis@ux1.cso.uiuc.edu (Joseph Angelis) Subject: Re: Linux Mud Date: 5 Oct 1994 23:54:21 GMT francis@VIOLET.uthscsa.edu (Scott Francis) writes: >Is there a mud developed for Linux and if so is it possible for me to get >the source or compiler version of it? Most of the Diku variants seem to work on my machine.. I'm using 1.1.0 of the kernel and the latest gcc. Haven't had any problems yet.. >Please respond to francis@violet.uthscsa.edu >Thanks is advance >Scott >////////////////////////////////////////////////////// >// // >// Scott Francis - UT Health Science Center // >// San Antonio, Texas // >// // >// e-mail: francis@violet.uthscsa.edu // >// sfrancis@janus1.cs.trinity.edu // >// // >// voice: (210)829-5501 // >// // >////////////////////////////////////////////////////// ------------------------------ From: sheneman@cs.uidaho.edu (Lucas James Sheneman) Crossposted-To: comp.os.linux.misc,comp.os.linux.help Subject: Help With LCD, Cirrus and XFree86 Date: 7 Oct 1994 02:03:36 GMT I just got XFree86-3.1 and installed it on my machine. I have a Cirrus-6440 chipset. I am trying to run the XF86-SVGA server. I managed to make the server happy by telling it I had a Cirrus-6420 chipset (like the docs say to do). When the server comes up, I get a nice, crisp, clear display on my LCD notebook, except for the fact that the top 50 pixels or so and the bottom 50 pixels of the screen are screwed up. The bottom 50 pixels of the display are identical to the top 50 pixels on the display...only shifted by about 100 pixels. My question is this: I have a dual-scan color LCD display and a cirrus 6440 chipset. Does it make sense for me to determine the video modes for my display as discussed in the text file VideoModes.txt? How do I treat an LCD display when trying to determine video modes and sync frequencies? Does my problem sound like a video timings problem or does it sound like a problem in which my server is mis-using VRAM? Specifically, here is my system: 1. Sager NP7500 Notebook/dual scan color LCD (640x480x256) 2. 1MB VRAM. 3. XFree86-3.1 Please e-mail me if you have any suggestions. Thanks. -Luke. -- +--------------------------------------------------------------------------+ | Luke Sheneman sheneman@cs.uidaho.edu | +--------------------------------------------------------------------------+ | Laboratory for Applied Logic, University of Idaho | | http://www.cs.uidaho.edu/lal/students/sheneman.dir/sheneman.html | +--------------------------------------------------------------------------+ ------------------------------ From: phil@zeus.fasttax.com (Phil Howard) Subject: Re: VESA and SVGAlib? Date: 6 Oct 1994 22:45:47 -0500 hhenson@inyanga.cs.wits.ac.za (Howard P. Henson) writes: >Andy Bailey (bailey9@muvms6.wvnet.edu) wrote: >: I have what might be a dumb question, about SVGAlib and video modes. I don't >: know diddly about programming graphics drivers, but here goes. >: With most DOS applications, namely graphics viewers, instead of specifying >: the drive specific to my card, I simply use the VESA driver, and voila, all my >: cardsmodes are recognized. Would this be possible for SVGAlib? I have overheard >: bits of conversation among Linux developers about avoiding making BIOS calls ( >: I guess to ensure portability to other processors). Is the case the same here? > The problem is not so much portability as the fact that VESA >drivers (and I asume you have a bios version) is 16 bit code and to run >it would require on to change CPU modes, and start thunking (16 -> 32 bit >conversions) etc. producing i) Slow access ii) Possible unstabilities on >the system which may land up with system crashes etc. VESA is 2 things: 1. A common BIOS programming interface. 2. A driver that understand your hardware. Linux has: 1. A non-BIOS programming interface. 2. Good drivers for documented hardware and some drivers for others, too. If your hardware is not supported by Linux: 1. Maybe someone is working on it now. 2. Call your vendor and ask for a "Linux driver". 3. Call your vendor and ask for full hardware documentation for driver development. 4. Sell your hardware to some DOS loser and get something better. -- /***** Phil Howard KA9WGN *********** How about universal JOBS? **************\ * Unix/Internet/Sys Admin Let's de-Foley-ate congress in 94 * * CLR/Fast-Tax Don't let Annie get your gun! * \***** phil@fasttax.com ************* Just say NO to CIX extortion ***********/ ------------------------------ From: ftlofaro@unlv.edu (Frank Lofaro) Subject: Re: Compiling progs using port I/O Date: Fri, 7 Oct 94 03:07:40 GMT In article <36hm55Ebnr@uni-erlangen.de> bon@lte.e-technik.uni-erlangen.de (Uwe Bonnes) writes: >Rob Janssen (rob@pe1chl.ampr.org) wrote: >> In <36bmo0$fmg@clarknet.clark.net> nardone@clark.net (Joe Nardone) writes: > > >> >Hey net-folks-- > >> >I'm trying to compile a program that uses the inb and outb >> >functions (macros, actually) but when it comes to link time >> >all my inb/outb calls are represented as unresolved references >> >to ___outb (or ___outcb) and ___inb... > >> >Am I missing a library, or a path to one? gcc -v looks like it's >> >looking in all the right places for library files... > >> >I'm running GCC 2.5.8 on Linux Kernel 1.1.50 w/ a 486dx2/66. > >> >Any help would be much appreciated- > >> You have to compile with optimization (-O2) > >Is there some explanation for that behaviour? >-- >Uwe Bonnes bon@lte.e-technik.uni-erlangen.de Yes, there is. extern inline is used instead of static inline in the function definitions. This is a Bad Thing. Without optimization, inline is ignored, thus extern inline becomes extern, which means it is not defined. Using static inline would mutate to static when not optimizing, which would still work (it would not be inline of course, but it wouldn't fail to compile, either). P.S. Using extern does not require -O2, just -O will work, Still, it would be nice to be able to use those functions even when you do not want to use optimization. ------------------------------ From: dehartog@kwetal.comcons.nl (Hans de Hartog) Subject: Re: Adaptec AHA-2940 PCI SCSI card support.... Date: 7 Oct 1994 13:10:23 -0000 Edward S Peschko (pesc0002@gold.tc.umn.edu) wrote: : hey all -- : Any plans (*please*) for developing support for the AHA-2940 PCI SCSI : board?? Yes, I've read somewhere that people are working on the PCI SCSI drivers Stay tuned to c.o.l.announce : Ed : (ie: This really sucks. I have a $5K system, with a 4MB VRAM card (and yes, : its diamond unfortunately) 4xPlextor, etc. etc... and its TOO NEW for linux! : If something doesn't come out soon, I guess it is just going to have to be : SCO...) Yup, that's life. There is a HOWTO-hardware (a.k.a. hardware compatibility list) for Linux. Better check that list before spending... -- _____________________________________________________________________________ Hans de Hartog, dehartog@comcons.nl, Voice: +31 348033100, Fax: +31 348033181 Committed Consultancy BV, Korenmolenlaan 1b, 3447 GG Woerden, The Netherlands Home: dehartog@kwetal.comcons.nl, Voice/Data: +31 838038560, CIS: 100121,3301 ------------------------------ From: hpa@ahab.eecs.nwu.edu (H. Peter Anvin) Subject: Re: ext2fs vs. Berkeley FFS Reply-To: hpa@nwu.edu (H. Peter Anvin) Date: Fri, 7 Oct 1994 04:11:17 GMT Followup to: <36nng2$4a1@babyblue.cs.yale.edu> By author: hstrong@eng1.uconn.edu (Hugh Strong) In newsgroup: comp.os.linux.development > > I know this doesn't sound very UNIXy, but how difficult would > it be to implement a filesystem with some sort of arbitrary > extended attributes like those in NTFS or the more limited ones > in HPFS? > Probably quite trivial (all you need is an attribute on a directory that marks it as "look like a file"). A file with multiple data streams (forks, to use Mac terminology) is nothing but a directory. However, you would have to teach ls, cp etc about this convention, which would be a pain. /hpa -- INTERNET: hpa@nwu.edu --- Allah'u'abha --- IBM MAIL: I0050052 at IBMMAIL HAM RADIO: N9ITP or SM4TKN FIDONET: 1:115/511 or 1:115/512 STORMNET: 181:294/1 or 181:294/101 "All sysadmins love logs." -- Me after deleting 87 Mb worth of log files ------------------------------ From: phl@cyways.com (Peter H. Lemieux) Crossposted-To: comp.os.linux.help,comp.os.linux.misc,comp.os.linux.admin Subject: Re: Telnet & ftp freeze! Date: 7 Oct 1994 04:26:30 GMT In article <3728nr$eb0@news.halcyon.com>, ralphs@halcyon.halcyon.com (Ralph Sims) says: > >Other things that run are Sendmail+IDA as a daemon, xntpd, and >CERN's web server. The ftp session definitely takes over the system. > Remember that FTP is running two simultaneous sessions with the other host, a data channel and a control channel. With only a standard two-wire modem, the line must be repeatedly turned around from TX to RX and back again. (Four wire, dedicated-line modems are pricey.) Services like news and the web have little upstream traffic, mostly downstream, since they have no control channel. Peter Dr. Peter H. Lemieux cyways, inc Voice: +1 (617) 924-7991 203 Arlington Street Fax: +1 (617) 926-8440 Watertown, MA 02172-2036 USA Internet: phl@cyways.com ------------------------------ From: dlj0@Lehigh.EDU (DAVID L. JOHNSON) Crossposted-To: comp.os.linux.misc Subject: Re: Beautifying Linux/Xfree Date: 7 Oct 1994 16:09:25 GMT In article <372tg0$1ai@huron.eel.ufl.edu>, acg@kzin.cen.ufl.edu (Alexandra Griffin) writes: >In article <1994Oct5.141142.773@muvms6>, >Andy Bailey wrote [in c.o.l.misc]: >>[...] >> >You may get some negative feedback from the die-hard functionality >over form crowd, but I'd say there's a lot of truth to what you're >saying-- cleaning up minor things like GUI look & feel *will* make the >X environment more appealing to a lot of people (maybe this is >unfortunate, but with all the "fluff" on a typical Mac/Windoze desktop >nowadays a lot of people expect this kind of thing...) We got LOTS of fluff, but no *one* set of it. That is sort of the main difference. You want standardized interfaces? Not likely. A vendor could provide one -- most of them do. But linux is a creature of the net, and there will not be only one GUI interface. > >Some ideas along this line that I've thought about: > >1) A mouse-driven tool for setting common X resource preferences would >be *very* helpful, even for experienced users (kind of a big project, >I know). I don't think this is too hard, and it is a good idea. >2) A better X file manager than what's currently out there (xfm & >xfilemanager are nice but not as easy to configure, easy to use, or >generally polished as one might like). Maybe something that provided >essentially the same functionality as Mouseless Commander (the >text-based Norton Commander clone), but with a mouse-driven GUI? (& >provisions for icons if desired, scrollbars on the dual file selection >lists, real pulldown menus-- leave in the command line at the bottom, >though!). This is arleady avialable. Try GREAT. Really. It takes a while to configure, and you should have Motif to get the best performance, but it has lots of options. > Well, you mentioned NextStep-- on second thought, something >similar to the wonderful NeXT Workspace Manager application would be >delightful to have. For those who have never seen it, this program >(in its Browser mode) presents a group of side-by-side vertical >directory listings, with each column representing a level of the >directory hiearchy... Y'know, the browser from Ghostview is similar to this -- not as advanced, though. GREAT's file manager is similar, as well -- though not the tree-structure. To each his own >3) Another idea from HP-VUE... this environment features a "console >bar" area at the bottom of the screen, containing buttons to switch >virtual desktops, invocation icons for commonly-used apps, small icons >for system functions (logging out...), and space for a clock, >calendar, Xload bargraph, & other stuff. Again, GREAT -- along with what you can do with .xinitrc, can do this. There is also another program out there that does things like this. Can't remember the name, but check them out as they show up. Much of your suggestions are really already available, and I don't see any interest in somehow standardizing them. Making such bells&whistles available is one thing, making them ubiquitous is another. Don't just assume that, if it isn't in slackware, it's not available. -- David L. Johnson dlj0@lehigh.edu or Department of Mathematics dlj0@chern.math.lehigh.edu Lehigh University 14 E. Packer Avenue (610) 758-3759 Bethlehem, PA 18015-3174 (610) 828-3708 ------------------------------ From: eric@aib.com (Eric Youngdale) Subject: Re: Shared Libs: working toward a permanent solution? Date: Fri, 7 Oct 1994 04:38:26 GMT In article , Richard Krehbiel wrote: >In article <36pe51$s0v@strauss.udel.edu> mike@strauss.udel.edu (Michael James Porter) writes: > >> In article , >> David Holland wrote: >> =>How about dynamically relocating the library when it's loaded - once. >> =>Then the address it appears at can be determined at run time; that way >> =>it cannot possibly conflict with any other libraries; the library >> =>loading mechanism would pick addresses so that doesn't happen. Then >> =>when other processes add it, it would appear at the same address in >> =>every process. >> >> >> This is a good idea > >I also like this idea; it does not require that library code be >position independent. There was a long thread about the actual >performance penalty of PIC; from that I gather that it's at least 3% >and possibly more. The performance penalty is not due to the library being relocated. It is because we lose one of the machine registers so that it can be used to point to the GOT table. The Intel architecture is painfully short of registers, so losing one gives a noticable performance impact. >> , but it limits the loading of a shareable library >> to program load time. I would like to see dlls be loadable at any >> time in the life of a process. > >I don't see why this is a problem; if it can be done at program load >time, it can certainly be done at run time. It only requires that the >load-and-relocate function (and perhaps the locate-symbolic-entry- >point functions) be available to executing processes. The only way to avoid the performance penalty is to modify the compiler to do something like half-pic. THis way the same pic operands are used so but are not referenced to the %ebx register - then you can free up the %ebx register. Next you need to make sure that the assembler generates sensible relocations for all of these operands, and then you can generate a shared library. Finally you can use some kind of utility to "fast-load" the shared library at some fixed address and perform the majority of the relocations and then write the file back out again. You will also need to modify the dynamic loader to recognize that the library has already been relocated and properly load it at the correct address - in this case the addresses would be chosen locally. The disadvantage is that with a system such as this the libraries would be painfully slow before they are "fast-loaded". If I were trying to improve performance of ELF libraries, I would probably try and do something along the lines of what I just outlined. The key is to avoid cumbersome build procedures and it simply must be as idiot-proof as possible. I want to reiterate that for now the concern is simply to get everything stable and debugged so that it works. At the moment there are some C++ issues that should be resolved, and then we will sit back and pound on it for a while to make sure that it all hangs together before we release it to the general public. Finally, I noticed earlier in this thread that someone said that it was the performance of linux that was so attractive, and that if we started doing this sort of thing (i.e. ELF), that they would switch to something like Solaris. I found this comment *extremely* amusing, as would anyone else who had ever used Solaris. In my new job I have occasion to work on a Solaris machine, and believe you me, Solaris is the butt of many rather rude jokes, all of them concerning poor performance. BTW - Solaris uses ELF, but the performance problems are apparenetly more to do with the kernel/filesystem rather than anything else - Unixware also uses ELF, and it is much faster. -Eric ------------------------------ From: martin@goofy.zdv.Uni-Mainz.DE (Christoph Martin) Crossposted-To: comp.windows.x.i386unix,comp.os.linux.admin Subject: Xfree 3.1 and SPEA MirageP64 (Linux) Date: 06 Oct 1994 16:11:53 GMT README.S3 (in XF86-3.1-doc.tar.gz) says: >1 - Supported hardware >---------------------- > > ... > >S3 864, 20C498 RAMDAC, ICS2595 Clockchip > SPEA MirageP64 2MB DRAM > > 8 and 15/16 bpp > > ClockChip "ICS2595" I tried this in my XF86Config file. The card is probed correctly as S3 864 with 20C486 RAMDAC. But the server can't set the clockchip. What is the problem? Is it the right ClockChip statement? Has anyone this card running with XFree 3.1? Christoph -- ============================================================================ Christoph Martin, Zentrum für Datenverarbeitung, Uni-Mainz, Germany Internet-Mail: Christoph.Martin@Uni-Mainz.DE Paper-Mail: C. Martin, Zentrum für Datenverarbeitung, Johannes-Gutenberg-Universität, 55099 Mainz, Germany Telefon: +49 6131 396316 -- ============================================================================ Christoph Martin, Zentrum für Datenverarbeitung, Uni-Mainz, Germany Internet-Mail: Christoph.Martin@Uni-Mainz.DE Paper-Mail: C. Martin, Zentrum für Datenverarbeitung, Johannes-Gutenberg-Universität, 55099 Mainz, Germany Telefon: +49 6131 396316 ------------------------------ From: ralphs@halcyon.halcyon.com (Ralph Sims) Crossposted-To: comp.os.linux.help,comp.os.linux.misc,comp.os.linux.admin Subject: Re: Telnet & ftp freeze! Date: 7 Oct 1994 14:06:44 GMT phl@cyways.com (Peter H. Lemieux) writes: >>CERN's web server. The ftp session definitely takes over the system. >Remember that FTP is running two simultaneous sessions with the other host, >a data channel and a control channel. With only a standard two-wire modem, Right, but introducing a 3000ms latency in the PPP session is not really acceptable. ------------------------------ From: killianr@beldin.sun.ac.za (Richelo Killian) Subject: LINUX Logical volumes Date: 7 Oct 1994 14:13:31 GMT I have a interesting question for all you LINUX gurus out there... Is it posible to create logigal volumes across drives and/or partitions and then mount a single filesystem on that volume? I know it can be done on HP-UX, but I want to do it on my LINUX box? Cheers Richelo Killian KILLIANR@TELKOM09.telkom.co.za ------------------------------ From: pats@mack.RT66.com (Patrick Spinler) Crossposted-To: comp.windows.x,comp.windows.x.motif,comp.windows.x.open-look,comp.windows.misc,comp.programming,comp.unix.bsd Subject: Freeware Motif/Xview/tty interface library ? Date: 07 Oct 1994 03:06:13 GMT Hi folks, I'm part of a group writing a set of administration tools for freeware unixes, and I'm looking for a freeware GUI interface library or class library that transparently supports or can be made to support several different interfaces, including Motif, Openlook, Xview, and tty (curses). MS-Win, Mac, et al are nifty extras but not necessary. Sorry for all the cross posting. Please email responses. If enough people are interested, I'll post a summary to comp.windows.x or comp.windows.misc. Thanks ! -- Pat ---- Patrick Spinler pats@rt66.com Re: flames - "May your cats grow hands" -- Patrick Spinler pats@rt66.com Re: flames - "May your cats grow hands" ------------------------------ ** 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 ****************************** .