From: Digestifier To: Linux-Development@senator-bedfellow.mit.edu Reply-To: Linux-Development@senator-bedfellow.mit.edu Date: Sat, 30 Oct 93 02:13:19 EDT Subject: Linux-Development Digest #198 Linux-Development Digest #198, Volume #1 Sat, 30 Oct 93 02:13:19 EDT Contents: Re: PCMCIA Xircom Ethernet driver? (Donald J. Becker) Re: Intel EE 16, register/progr. docu, driver (Donald J. Becker) Re: Slowness in scsi disk access (Keith Smith) Re: Andrew File System (David Willmore) Re: Slowness in scsi disk access (Gerhard Fuernkranz) which (proc-) ps output shows the real memory usage? (Harald Koenig) Re: ugly name for core dumps (core.imagename) -> patch for "img.core" (Johan Myreen) Linux as a router?! (Jay Allen) Re: What's wrong with a DOS to Linux disk access? (Dave Truckenmiller) Linux/X386 on Dell Optiplex with TSENG W32i chip???? (Mr Andrew Simpson) Re: What's wrong with a DOS to Linux disk access? (Alan Cox) PCI-Bus on Linux ?? (Thilo W. Pannen) Re: What's wrong with a DOS to Linux disk access? (David Barr) Re: What's wrong with a DOS to Linux disk access? (Colin Owen Rafferty) Re: PCMCIA Xircom Ethernet driver? (Mark Eichin) Dos discrimination, Politically Correctness in the Linux community. (Just a fellow traveller...) Re: What's wrong with a DOS to Linux disk access? (khockenb@vaxc.stevens-tech.edu) ---------------------------------------------------------------------------- From: becker@super.org (Donald J. Becker) Subject: Re: PCMCIA Xircom Ethernet driver? Date: Thu, 28 Oct 1993 16:04:42 GMT In article , Kwun Han wrote: > I am wondering if there are any drivers available for the >Xircom PCMCIA ethernet card around? or does the current ethernet >driver support that (I don't think it will, but I will ask anyways). >If someone have any success using that card in their notebooks, I >would like to know how.. thanks a lot!! Xircom is known for not releasing programming information. Even if they change their policy you are unlikely to see a driver from me. -- Donald Becker becker@super.org IDA Supercomputing Research Center 17100 Science Drive, Bowie MD 20715 301-805-7482 ------------------------------ From: becker@super.org (Donald J. Becker) Subject: Re: Intel EE 16, register/progr. docu, driver Date: Thu, 28 Oct 1993 23:31:28 GMT In article <2alfe4$d9i@urmel.informatik.rwth-aachen.de>, wrote: >Who has documentation (registers/programming) at hand for Intels >EthernetExpress 16 card? (Intel EE 16 for Thin/Thick Ethernet = >BNC or tranceiver). I've read that there is an alpha release driver >avaiable - it is in the SLS package? (if yes, where?) I heard rumors that you must know the author personally, and perhaps even bestow gifts upon him, before he'll tell you how to become an alpha tester. Even then he expect you to send in good bug reports and fixes. I've also heard that he worked from an old 82586 databook, and later read the Crynwr driver rather than trying any longer to find the right people at Intel. I've checked the SLS package, and I haven't found the alpha EtherExpress driver there. Bummer, eh? I guess we'll have to look elsewhere for it... -- Donald Becker becker@super.org IDA Supercomputing Research Center 17100 Science Drive, Bowie MD 20715 301-805-7482 ------------------------------ From: keith@ksmith.com (Keith Smith) Subject: Re: Slowness in scsi disk access Date: Thu, 28 Oct 93 02:50:56 GMT In article <5144@blue.cis.pitt.edu>, Bill Broadley wrote: >Here's my theory: [ Rotational latency in the code problem description ] Neat. Except all the modern SCSI disks have at least 32K of on board cache, so the next request should come from the DRIVES cache. The 1G's I'm getting now come with 256K of cache on board. Rotational latency problems should be minimal / not a factor on a drive like this as it will ALWAYS be buffering a whole track. IDE will generally _always_ beat scsi in a raw thruput test, at the expense of about 4 times the CPU overhead. -- Keith Smith keith@ksmith.com 5719 Archer Rd. Digital Designs BBS 1-919-423-4216 Hope Mills, NC 28348-2201 Somewhere in the Styx of North Carolina ... ------------------------------ From: willmore@iastate.edu (David Willmore) Subject: Re: Andrew File System Date: Fri, 29 Oct 1993 04:32:55 GMT m147150@urz-mail.urz.uni-heidelberg.de (Markus Nullmeier) writes: >there a couple of months ago John F. Carr of MIT summarized a evaluation >of AFS with something like >"a) AFS filesystems are not Unix filesystems > b) the implementation quality of AFS is poor" Hey, if someone at MIT calls an implementation poor, it's got to suck. People at MIT know about poor quality implementations. Just look at kerberos, zephyr, moira, etc... (Just a little good natured razzing for our friends over at MIT. :) Cheers, David -- ___________________________________________________________________________ willmore@iastate.edu | "Death before dishonor" | "Better dead than greek" | David Willmore | "Ever noticed how much they look like orchids? Lovely!" | =========================================================================== ------------------------------ From: fuer@nessie.gud.siemens.co.at (Gerhard Fuernkranz) Subject: Re: Slowness in scsi disk access Date: Fri, 29 Oct 1993 14:24:50 GMT Eric Youngdale (eric@tantalus.nrl.navy.mil) wrote: : : I do not view it as a replacement, I view it as a more specialized tool : that would respond dynamically to the system requirements. In my mind both : would continue to exist, but they would be doing somewhat different things. : The bdflush (fsflush on our SVr4 machine) would not write all dirty buffers, : but would endeavor to keep the top of the free list supplied with clean : buffers. Thus it would not always write everything dirty in the buffer cache, : but only perhaps the first 500 dirty buffers. The update task could : continue to sync every 30 seconds, and would serve a slightly different : purpose. Eric, are you *really* sure, that SVR4 fsflush does *not* write all dirty buffers back to disk, even if you 1) wait long enough (eventually much more than 30 sec. when enough mem/buffers are free) and 2) you don't produce new dirty pages in the meantime ? There are two reasons, why I hate sync every 30 seconds: 1) It produces a load peak every 30 seconds (has already been discussed in this thread). 2) It writes temporary files (e.g. during compilation) to disk, even if the whole file fits into the buffer cache and the file will be removed 1 minute later. Bdflush with an LRU strategy would write other (old) blocks to disk instead and the temp file would no longer exist, when its buffers were old enough (This example assumes, that you have a machine with lots of memory). Gerhard -- Gerhard Fuernkranz SIEMENS Austria Corp. Email: fuer@siemens.co.at Gudrunstr. 11 Phone: +43-1-60171-5716 A-1100 Wien ------------------------------ From: koenig@ceres.tat.physik.uni-tuebingen.de (Harald Koenig) Subject: which (proc-) ps output shows the real memory usage? Date: 29 Oct 93 15:20:53 GMT I'm using 0.99pl13n at the moment with a recent procps (0.8 ?). I ever wondered why 'ps ux' and 'ps mx' show different memory statistics. Here an example: # ps ux USER PID %CPU %MEM SIZE RSS TTY STAT START TIME COMMAND root 1 0.0 1.3 44 200 ? S 23:58 0:00 init auto root 259 0.0 2.4 232 360 ? S 00:08 0:26 xntpd root 2073 5.0 1755.2 1735 261676 ? S 08:56 0:57 X :0 # ps mx PID TTY MAJFLT MINFLT TRS DRS SIZE SWAP RSS SHRD LIB DT COMMAND 1 ? 66 37 8 88 372 0 372 276 320 11 init auto 259 ? 52 40 116 148 544 0 544 292 324 11 xntpd 2073 ? 275 2982 520 1720 2852 0 2852 380 1444 208 X :0 Note the different numbers for SIZE and RSS. What happens with RSS for my X server (this is XFree86-2.0 S3 server which maps the 1MB video frame buffer at address 0x7c00000 via mmap(2))? When I disable the frame buffer mapping, 'ps ux' shows a reasonable RSS number. Any explanation and/or fix? Thanks, Harald -- Harald Koenig, Inst.f.Theoret.Astrophysik (koenig@tat.physik.uni-tuebingen.de) All SCSI disks will from now on be required to send an email notice 24 hours prior to complete hardware failure! ------------------------------ From: jem@snakemail.hut.fi (Johan Myreen) Subject: Re: ugly name for core dumps (core.imagename) -> patch for "img.core" Date: 29 Oct 93 15:54:57 GMT In article <2aiqvd$1n1@smurf.sub.org> urlichs@smurf.sub.org (Matthias Urlichs) writes: >True. I haven't seen any objection to naming the corefile of "foo" >"foo.core". Quite a lot of people are still using the Minix file system, with a maximum file name length of 14. -- Johan Myreen jem@cs.hut.fi 60 11'55" N, 24 53'30" E ------------------------------ From: hgja@acacia.cs.pdx.edu (Jay Allen) Subject: Linux as a router?! Date: 29 Oct 1993 19:30:29 GMT Has anyone actually got a eth0<-->eth1 routing settup under Linux? All I did was add a second struct to Space.c and made a new kernel, and it WORKS!! Well most of the time... On occaision the TCP clients (MAC and Windows) will hang while logged on to my router/host. The console will be hung and nothing but a poweroff will bring it back! If anyone is using Linux as a router (sucessfully!) please post me, or email me directly at: hgja@odin.cc.pdx.edu Regards, Jay Allen ------------------------------ From: trucken@giga.cs.umn.edu (Dave Truckenmiller) Subject: Re: What's wrong with a DOS to Linux disk access? Date: Fri, 29 Oct 1993 19:50:23 GMT >That is not what Christian is asking. He wants to be able to grab a >Linux file when the machine is running in DOS/Windows mode. Right >now, the only way to do this is to reboot in Linux mode, log in, do >the copy, and the reboot in DOS mode. >> I would tend to think that a DOS reader would be of limited utility, >> at best. Am I missing something? >Yes. I created a text file in emacs on the Linux disk, and now, when >I'm running Quark under Windows, I want to import it. How do I get it? Yes, this is EASY. Mount your DOS partition in Linux. Make it your HOME directory. :-) Edit any number of files with emacs. Reboot to DOS and your files are still in your DOS directory. I do this ALL THE TIME! This is a silly thread. There is no 'unix filesystem', there are several filesystems: minux, ext, ext2, xiaf. By thinking ahead only slightly, you and arrange to have all the files you share between Linux and DOS always live on the DOS partition. (Use ln -s if you want!) I think you can even mount DOS so that cr -- cr/lf converstion is automatic. -Dave -- --- Dave Truckenmiller (trucken@cs.umn.edu) [ ASCII picture ] Linux, Linux, Linux, Linux, Linux, Linux. [ under development ] ------------------------------ From: uclyams@ucl.ac.uk (Mr Andrew Simpson) Subject: Linux/X386 on Dell Optiplex with TSENG W32i chip???? Date: 29 Oct 1993 13:56:25 -0500 Does anybody know if it possible to run Linux/x386 on a Dell Optiplex machine; Dell have told me that their local bus video card uses the Tseng w32i chip; is this compatible with other supported chips, or is this board supported by Linux/x386. Any help would be gratefully received. Andrew Simpson andrew@phon.ucl.ac.uk ------------------------------ From: iiitac@swan.pyr (Alan Cox) Subject: Re: What's wrong with a DOS to Linux disk access? Date: Fri, 29 Oct 1993 18:12:23 GMT In article craffert@nostril.lehman.com (Colin Owen Rafferty) writes: >Yes. I created a text file in emacs on the Linux disk, and now, when >I'm running Quark under Windows, I want to import it. How do I get it? With difficulty. If someone was sick enough they could write a dos file redirector that used the BIOS to do disk i/o and itself implemented the relevant linux fs you were using (minix/ext2 or whatever). From a programming point of view you would have to be really deranged to do this. Much better to think of it in time before you switched OS and rebooted! Alan iiitac@pyr.swan.ac.uk ------------------------------ Crossposted-To: comp.os.linux From: pannen@gmd.de (Thilo W. Pannen) Subject: PCI-Bus on Linux ?? Date: Fri, 29 Oct 1993 15:53:33 GMT Hi out there, is there any work in progress on porting Linux to recently released Intel PCI-bus-based systems ? Thanks, Thilo -- Thilo W. Pannen, E-mail: pannen@gmd.de German National Research Center for Computer Science Phone: (+49) 2241-14-2547, Fax: (+49) 2241-14-2105 ------------------------------ From: barr@pop.psu.edu (David Barr) Subject: Re: What's wrong with a DOS to Linux disk access? Date: 29 Oct 1993 20:18:15 GMT In article , Dave Truckenmiller wrote: >>Yes. I created a text file in emacs on the Linux disk, and now, when >>I'm running Quark under Windows, I want to import it. How do I get it? > >Yes, this is EASY. Mount your DOS partition in Linux. It's hard to believe so many people responding to this thread can be this thick, and even after it being re-explained. What he's asking is how to access the Linux filesystem FROM DOS! I don't know of a way, you'd have to write a Linux filesystem driver that runs under DOS, and I don't know that one exists. --Dave -- System Administrator, Penn State Population Research Institute "Diplomacy is the art of saying 'nice doggy' until you can find a rock." - Will Rogers ------------------------------ From: craffert@nostril.lehman.com (Colin Owen Rafferty) Subject: Re: What's wrong with a DOS to Linux disk access? Date: Fri, 29 Oct 1993 13:27:01 GMT >>>>> In article <2aq2ju$h9i@crl.crl.com>, rwatkins@crl.com (Ronald Watkins) writes: > Christian Holtje (choltje@ux1.cso.uiuc.edu) wrote: > :I'm not very good at programming, but I'd be willing to try helping: > :I think some one should write a program in DOS to access the linux > :partitions. > : Everytime I've seen someone ask about accessing Linux disks from DOS, > : they're smashed down, with few reasonable responses, even though this > : is a reasonable question. > Um, I'm not sure why this is necessary? I can transfer files both ways > without a problem from Linux. I just mount the MSDOS file system as > read/write, and cp away. Blazing fast and easy as anything. That is not what Christian is asking. He wants to be able to grab a Linux file when the machine is running in DOS/Windows mode. Right now, the only way to do this is to reboot in Linux mode, log in, do the copy, and the reboot in DOS mode. > I would tend to think that a DOS reader would be of limited utility, > at best. Am I missing something? Yes. I created a text file in emacs on the Linux disk, and now, when I'm running Quark under Windows, I want to import it. How do I get it? -- Colin Rafferty, Lehman Brothers pgp print = 91FED077 BD5588B6 30B372D2 F9172162 Don't know what pgp is? Ask me! ------------------------------ From: eichin@cygnus.com (Mark Eichin) Subject: Re: PCMCIA Xircom Ethernet driver? Date: 29 Oct 1993 13:17:54 -0400 >> Hmm... so are there *any* PCMCIA ethernet card supported by Linux? I Socket Communications makes a NE2000-compatible PCMCIA ethernet card; I've seen it work under older versions of linux. It just needs a bit of code to "enable" it, and then the normal NE2000 driver autoprobe takes care of the rest. Enabler code exists for the 82356 (Intel chip, used in Dell among others) but I don't have time to make it available (nor does the guy who wrote it.) It's not that clean anyway; hopefully Barry Jaspan's PCMCIA drivers will work well enough to suffice. I am working on Databook TCIC drivers (Databook chip, used in TwinHead/Compudyne machines among others) and will release what I end up with, but haven't had time to do much with it. Probably some time in December I will. _Mark_ MIT Student Information Processing Board Cygnus Support ------------------------------ From: jedubins@unix.amherst.edu (Just a fellow traveller...) Subject: Dos discrimination, Politically Correctness in the Linux community. Date: 29 Oct 1993 20:15:10 -0400 Alan Cox (iiitac@swan.pyr) wrote: : In article craffert@nostril.lehman.com (Colin Owen Rafferty) writes: : >Yes. I created a text file in emacs on the Linux disk, and now, when : >I'm running Quark under Windows, I want to import it. How do I get it? : With difficulty. If someone was sick enough they could write a dos : file redirector that used the BIOS to do disk i/o and itself implemented : the relevant linux fs you were using (minix/ext2 or whatever). From a : programming point of view you would have to be really deranged to do this. : Much better to think of it in time before you switched OS and rebooted! : Alan : iiitac@pyr.swan.ac.uk I really don't want to start a flamewar here, because it seems to be such a senseless waste of human life, but.... I've been using and programming on different platforms and in different languages for a long time. It's not like I just started using and programming computers yesterday. I've always realized how important it was to know that the computer was supposed to be my tool, and not the other way around. If it's useful, and saves time under a general situation or otherwise, why not implement it? The reaction to this situation reminds me of the old MS joke: Q: How many MS employees does it take to screw in a light bulb? A: None. MS redefines darkness to be the industry standard. Now I don't want to throw any names out or anything, but who in this community has taken that attitude with regards to this problem, and who has not? :) I would've thought it would be morally repugnant for most Linux developers to realize they had the same ways of approaching problems as MS employees. Hear that MS? You've got a bunch of recruits ripe for the picking... :) ------------------------------ From: khockenb@vaxc.stevens-tech.edu Subject: Re: What's wrong with a DOS to Linux disk access? Date: Sat, 30 Oct 1993 01:27:53 GMT In article <1993Oct29.181223.14981@swan.pyr>, iiitac@swan.pyr (Alan Cox) writes: > In article craffert@nostril.lehman.com (Colin Owen Rafferty) writes: >>Yes. I created a text file in emacs on the Linux disk, and now, when >>I'm running Quark under Windows, I want to import it. How do I get it? > With difficulty. If someone was sick enough they could write a dos > file redirector that used the BIOS to do disk i/o and itself implemented > the relevant linux fs you were using (minix/ext2 or whatever). From a > programming point of view you would have to be really deranged to do this. > Much better to think of it in time before you switched OS and rebooted! A simpler idea (if someone wants to write it) is to make a DOS equivalent of mtools, but in reverse. That is, a set of DOS programs like "lls" (linux ls), "lcp" (linux copy), etc. that can access a linux partition an read the files there. In fact, I stumbled across a partial implementation of something like this. There's a set of programs to read a minix-floppy from DOS on the simtel mirrors, in the dskutl dir. As I recall, I got a copy from oak.oakland.edu:/pub/msdos/dskutl/mindos11.zip. I didn't look too close at it though, because I don't use minix-fs anymore. -Kurt Hockenbury ------------------------------ ** 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 ******************************