Received: from SOUTH-STATION-ANNEX.MIT.EDU by po7.MIT.EDU (5.61/4.7) id AA26398; Sat, 16 Dec 95 22:15:18 EST Received: from SENATOR-BEDFELLOW.MIT.EDU by MIT.EDU with SMTP id AA19172; Sat, 16 Dec 95 22:15:17 EST Received: (from root@localhost) by senator-bedfellow.MIT.EDU (8.6.12/2.3JIK) id WAA29439 for Linux-Development-System-Dist@senator-bedfellow.mit.edu; Sat, 16 Dec 1995 22:13:43 -0500 Message-Id: <199512170313.WAA29439@senator-bedfellow.MIT.EDU> From: Digestifier To: Linux-Development-System@senator-bedfellow.MIT.EDU Reply-To: Linux-Development-System@senator-bedfellow.MIT.EDU Date: Sat, 16 Dec 95 22:13:40 EST Subject: Linux-Development-System Digest #126 Linux-Development-System Digest #126, Volume #2 Sat, 16 Dec 95 22:13:40 EST Contents: Re: Kernel protection against device drivers! (Joe Nardone) Re: ISDN and CAPI (Holger Petersen) Re: Are libc functions MT-safe? (Scott Schwartz) Re: Mounting a mac file-system under linux? (Prof R Boswell) Re: Linux has poor memory management? (John Burton) Re: tpqic02: LAST call for help (Peter Much) Kernel Change Summary 1.3.47 (Michael Elizabeth Chastain) Re: Linux has poor memory management? (Nick Kralevich) Re: Kernel resets IDE HD -- Normal? (Michael Faurot) AutoMounter (Mark Carey) Re: 'make' fails with 1.3.47? (Matthew Cummings) Re: Can Linux support multiprocessors INTEL CPU? (Andrew Mileski) Re: Does Linux support (or will) RAID 5 drives and/or Hot plugged drives ? (Andrew Mileski) Re: AMD DX4-120 (Derek Wildstar) Re: 1.3.45 kernel memory usage (Jason Mcmullan) Always IN-2000 SCSI card (Blue Angel) Re: Linux ISO 7816 emulator (was: New Voyager) (M.Buchenrieder) [Kernel Module] Upgrade To ELF ---------------------------------------------------------------------------- From: joe@access4.digex.net (Joe Nardone) Subject: Re: Kernel protection against device drivers! Date: 12 Dec 1995 21:49:31 GMT Aurel Balmosan (aurel@xylo.owl.de) wrote: : Hello, : Is there a project which developes a kernel protection against devices drivers. : [snip] : : Aurel Balmosan Yeah, they're called the 1.2 kernels. Seriously, how would you implement such a thing? It would be very difficult to impose those kinds of restrictions on device drivers, since they must run at a kernel-level and have that open access to the system and its hardware. That's kind of the point. If a driver misbehaves, it means it's probably still being developed, which means it's probably in a 1.3 kernel, which means you are on your own. Joe : : ---------------------------------------------------------------- : Aurel Balmosan aurel@xylo.owl.de : -- : : ---------------------------------------------------------------- : Aurel Balmosan aurel@xylo.owl.de -- =------------------------------------------------------------------= Joe Nardone "Now that I have my coffee, I'm ready to watch radar." =====BEGIN PGP PUBLIC KEY BLOCK===== Version: 2.6.2 mQBtAzCg5QoAAAEDANvEmc4OAT8wntnCv8UPjyD2kohKVGa1CU21PLpYljzl3AE7 pldpNNscKSoE//aVMKYBO+tRFTTsPNi0qZ6CDlEBXTgJAiszoHh1qCzZyYj7FTBL YoUr+3sk9rU8Pma6AQAFEbQiSm9lIE5hcmRvbmUgPGpvZUBhY2Nlc3MuZGlnZXgu bmV0Pg== =1ocE =====END PGP PUBLIC KEY BLOCK===== ------------------------------ From: hp@kbbs.org (Holger Petersen) Subject: Re: ISDN and CAPI Date: Sat, 16 Dec 1995 17:18:45 GMT honore@antok.news.univ-lille1.fr (Honore B.-LAMIH) writes: >Hi, >I'm looking for software (shareware or public domain >preferred) that uses the CAPI 2.0 interface under >Linux. pourquoi? [why another non-standard?] Is there any Unix-Software using any version of CAPI? Why not treat the ISDN-Line as just that: Another Inter- face which looks to unix just like any ohter device? Depending on the system at the other end treat it like a TCP/IP-Connection or a normal getty-driven modem? amicalement, Holger ------------------------------ From: schwartz@galapagos.cse.psu.edu (Scott Schwartz) Subject: Re: Are libc functions MT-safe? Date: 16 Dec 1995 06:02:14 GMT grib@cs.utexas.edu (William S. Gribble) writes: | My understanding is that clone() in *not* intended as a user interface to | threads, just as a thread-implementor's interface. So you can forget | about it unless you want to reimplement the POSIX or another thread | API. That's a pity. Plan 9's rfork(), similar to clone(), is a beautifully simple and effective way to do things, much nicer than pthreads. The idea is simple and clean: the OS has one (1) way to create a new process (== a thread of control.) fork() is simply a wrapper for rfork() with the right argument. If clone() were fully implemented in the Plan 9 style I think it would be very popular. ------------------------------ From: chem211@cantua.canterbury.ac.nz (Prof R Boswell) Subject: Re: Mounting a mac file-system under linux? Date: 16 Dec 1995 17:51:18 GMT Aaron Mchugh (ez048853@bullwinkle.ucdavis.edu) wrote: : Hi, I'm wondering if there are any apps to mount or read a mac : floppy on a linux system From the LSM: Begin2 Title = Macintosh HFS Access Tool Version = 0.36 Desc1 = This program allows non-Macintosh hosts to read Desc2 = Macintosh HFS floppy disks, CD-ROMS and hard Desc3 = drives Desc4 = Desc5 = Author = Craig Southeren AuthorEmail = Maintainer = Craig Southeren MaintEmail = Site1 = sunsite.unc.edu Path1 = pub/Linux/utils/disk-management File1 = hfs0_36.tgz FileSize1 = 57361 bytes -- +-----------------------------------------+-----------------------------+ | Ross Boswell | Email drb@chmeds.ac.nz | | Christchurch School of Medicine | FAX +64 3 364 0009 | | PO Box 4345, Christchurch, NEW ZEALAND | Phone +64 3 364 0590 | +-----------------------------------------+-----------------------------+ ------------------------------ From: john@ns.gats.hampton.va.us (John Burton) Subject: Re: Linux has poor memory management? Date: 15 Dec 1995 14:33:26 GMT In article albert@krakatoa.ccs.neu.edu (Albert Cahalan) writes: [...stuff deleted...] Someone else (tossed in the bit bucket by @%#*$%^&& emacs) wrote: > It says that RAM is a lot faster than the hard drive. > If you want better performance, add more RAM. No no no! This really is a problem specific to Linux. You missed the part about 'positively feeds back'. The complaint is not that swap space is slow (of course it is) but that a swapped process can never leave swap under the right conditions. No joke. To demonstrate: write a program that mallocs most of your RAM, say 6 MB on an 8MB machine. The program must (randomly?) touch all the pages over and over again. Just touch them, to keep the program fast. Now run this program in the background. Wait a bit, then run a second copy. When one process needs a page, the other one gets to run. The running process dirties its pages, so they will not get swapped (ok, paged) out. The process that needed a page gets one, but also loses a page. This causes the needy process to take another page fault, letting the pig run again and dirty more pages. The kswap patches added to 1.3.xx indirectly helped, but they did not solve the real problem. They did reduce the symptoms a bit though. Okay, sounds like Linux memory management has a problem, but what is the "real problem", and how do we solve it? You indicate that the kswap patches helped, but did not solve the "real problem" so it sounds like you have a clear understanding of what the real problem is. I don't, but would like to...could you please explain what the real problem is here ? John -- John Burton GATS, Inc. j.c.burton@gats.hampton.va.us 28 Research Drive j.c.burton@larc.nasa.gov Hampton, VA 23666 (804) 865-7491 (voice) (804) 865-1021 (fax) ------------------------------ Crossposted-To: comp.unix.bsd.freebsd.misc From: peter@citylink.dinoex.sub.org (Peter Much) Subject: Re: tpqic02: LAST call for help Date: Sat, 16 Dec 1995 08:00:35 GMT In article <4anj90$mqa@uriah.heep.sax.de>, J Wunsch wrote: >mt status for QIC-02 drives is still unsupported. Anyway, let's face >it: you don't really need it. :) It's mostly useless, since it would tell what the driver thinks about the tape, which is not necessary what is happening with the tape. On (earlier) Linux, it simply tells nonsense. >All other commands should work, as long as the underlying hardware >does support it. (It doesn't support block seeks, nor backward file You're an optimist.;) I had to change the wt.c file a bit to get it working in a reliable way: 1. The "mt fsf" command returns immediately, while the tape is winding. 2. The "dd" command reports an error after waiting 60 secs. on the winding, i.e. blocked, device. This together makes the tape drive quite unuseable for reliable com- pressed multivolume tar-archives. (tar cannot do that, one has to split the input stream into some-MB-chunks and compress the chunks separately - which is wiser anyway, since a broken tar file can be recovered more easily than a broken gzip file. To later collect the chunks back from tape, they need some headers. This gives mt fsf/dd/mt fsf/dd/... access sequences when appending, searching, verifying, or after an intermediate retension.) Since the wt.c file has excellent documentation in it, it is quite easy to change 1. The original behaviour was designed intentionally, i have no idea why. I don't know yet what's going on with "dd". Peter -- Write to: Peter Much * Koelnische Str. 22 * D-34117 Kassel * +49-561-774961 peter@citylink.dinoex.sub.org * much@hrz.uni-kassel.de ------------------------------ From: mec@treflan.shout.net (Michael Elizabeth Chastain) Subject: Kernel Change Summary 1.3.47 Date: 13 Dec 1995 19:59:12 GMT Reply-To: mec@duracef.shout.net Kernel Change Summary Linux 1.3.47 (5077 lines) Wed 13 Dec 1995 Documentation/Configure.help: many revisions (680 lines) [Axel Boldt]. drivers/char/tty_io.c: restrict vt switching via ioctl [grif@cs.ucr.edu]. drivers/char/vt.c: restrict vt switching via ioctl [grif@cs.ucr.edu]. drivers/net/arcnet.c: IRQ is always printed as a decimal, instead of a hex number. Replaced most remaining printk's with BUGMSG macros. Moved variables for VERIFY_ACK from the Outgoing struct into the lp struct; Outgoing is really for packet-splitting info only. Simplified INTMASK handling using a memory variable to keep a "running total". Tracked down more (maybe the last?) "missed IRQ" messages caused by TX-done IRQ's occurring WHILE arcnet??_send_packet was in progress. These were pretty obscure and mostly harmless. Made an actual non-ALPHA release of the thing. drivers/net/net_init.c: support alias devices (not enabled yet). fs/super.c: fix memory allocation bug in 'add_vfsmnt'. include/asm-i386/pgtable.h: page-granular invalidates for i486 and up. include/linux/net_alias.h: alias devices. include/linux/resource.h: change setpriority range from [-99,14] to [-20,20]. net/core/net_alias.c: alias devices. scripts/Configure: help text support [Axel Boldt]. Michael Chastain mec@duracef.shout.net Check out my burgeoning trace-and-replay debugger: ftp://tsx-11.mit.edu/pub/linux/sources/usr.bin/mec-0.3.tar.gz ------------------------------ From: nickkral@parker.EECS.Berkeley.EDU (Nick Kralevich) Subject: Re: Linux has poor memory management? Date: 15 Dec 1995 23:54:14 GMT This isn't a problem with Linux in specific -- this is a problem with all computer systems. In all of operating systems, there is a fundamental tradeoff between fairness and throughput. The two have to be balenced. For example, let's look at a scheduler that has to run a set of jobs. We are trying to optimize to minimize the number of seconds per job (seconds/job). Let's assume we know, in advance, how long each program is going to run. Also, let's assume that the context switch time is 1 seconds (yes, I know this isn't realistic). Then we might have the following 5 programs to run: program time -------------------------------- A 30 sec B 45 sec C 60 sec D 75 sec The optimal scheduling for this is A-B-C-D, which means that job A ends at 30 sec, job B ends at 76 sec, job C ends at 137 sec, and job D ends at 213 sec (on a uniprocessor). secs/job rating of: secs/job = (30+76+137+213)/4 = 114 sec/job Any other combination would have a higher sec/job rating. Now, if we allow other processes to start while these processes are running, then it is possible that D will get "starved" (i.e., never given any CPU time). So this scheme is unfair to long jobs, but gives optimal sec/job. Now, if we try to make it fair to longer jobs, by introducing somthing like time-slicing or multitasking, then we introduce a lot of overhead which will lower our throughput. Also, the different scheduling algorithm will raise our sec/job measure. For example, let's task switch every 5 seconds, and let each program have equal chance of getting run. Then the execution order is: A-B-C-D repeated 6 times B-C-D repeated 3 times C-D repeated 3 times D repeated 3 times "A" will end at time 126, "B" at time 186, "C" at time 228, and "D" at time 252. sec/job = 792/4 = 198 sec/job Using this scheduling algorithm, D will never get starved, but the drop in throughput is very noticable. Compounding this problem is the fact that the operating system doesn't know how long the programs are going to run. Anyway, the point of this example was just to point out somthing: Fairness and throughput are not compatiable. The fairest operating systems will usually be the ones that are least efficient. The most efficient operating systems will be the least fair. It is almost impossible to get around this problem, short of adding more CPUs. So you have to make a tradeoff between throughput and fairness. Take care, -- Nick Kralevich nickkral@cory.eecs.berkeley.edu ------------------------------ From: mfaurot@phzzzt.atww.org (Michael Faurot) Subject: Re: Kernel resets IDE HD -- Normal? Date: Thu, 14 Dec 1995 02:43:11 GMT Jeff Garretson (jeffgarr@cac.washington.edu) wrote: : The following is from /var/log/kernel: : : -------- : Dec 9 14:22:50 marley kernel: hda: read_intr: status=0x59 { DriveReady : SeekComplete DataRequest Error } : Dec 9 14:22:50 marley kernel: hda: read_intr: error=0x04 { : DriveStatusError } [...] : It's happened three times now, the first time was Nov. 24. I'm using : kernel version 1.3.28, and the drive is a 1.2 GB Western Digital Caviar. : I'm using the EIDE controller built into my Acer (AP-5C) motherboard. I believe I'm having a similar problem with a Maxtor 71260 (1.2G Drive). Here's an example from my /var/adm/messages file: Dec 13 18:19:56 phzzzt kernel: hdb: DRDY error: status=0x59 { DriveReady SeekComplete DataRequest Error } Dec 13 18:19:56 phzzzt kernel: hdb: DRDY error: error=0x10 { SectorIdNotFound }, LBAsect=105739, sector=311332 Thinking this was a problem with the drive, I called Maxtor and they graciously sent me a brand new drive. The above clip is from the new drive, just installed within the last several hours. At this point, I don't think there's anything wrong with the drive. Oddly enough, of the two IDE drives I have, these errors only seem to come up on the Maxtor. My other IDE, a Conner CFS420A (406M) never seems to burp at all. My kernel version is 1.2.13 with the IP masquerading patches installed. CPU is a 486-DX2/66, 24M of RAM with a DTC VLB IDE host adapter. -- +----------------+-------------------------+---------------------------------+ | Michael Faurot | mfaurot@phzzzt.atww.org | I programmed COBOL in college. | | ------- ------ | ------- ------ ---- --- | But I never compiled . . . :-) | +----------------+-------------------------+---------------------------------+ ------------------------------ From: Mark Carey Subject: AutoMounter Date: Thu, 14 Dec 1995 20:13:29 +0000 Is anyone developing an AutoMounter for linux? Or has it already been done? Thanks, Mark. -- -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Mark Carey, mark@dodgy.hiway.co.uk, mark_carey@mentorg.com. -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- ------------------------------ From: cummings@stingray.net (Matthew Cummings) Subject: Re: 'make' fails with 1.3.47? Date: 16 Dec 1995 18:25:11 GMT In article <4as4hp$2pnu@news-s01.ny.us.ibm.net>, sbeer@ibm.net writes: >BTW, when make is force-fed the top Makefile (-f switch) it gets stuck >deeper down looking for other rules (you can tell I'm not a 'make' >expert). I wonder if you upgraded the libc stuff and didn't read the release notes for that upgrade? If so get them because it has the patch you need to recompile make and have it work correctly. -- Internet: cummings@stingray.net ------------------------------ From: dmtech@magi.com (Andrew Mileski) Subject: Re: Can Linux support multiprocessors INTEL CPU? Date: 17 Dec 1995 02:05:01 GMT Sergei Viznyuk (sviznyuk@magnus.acs.ohio-state.edu) wrote: > In article <4aprpq$g91@nuscc.nus.sg>, Lee Yew Nam wrote: > >[ Article crossposted from comp.os.linux.hardware ] > >[ Author was Lee Yew Nam ] > >[ Posted on 14 Dec 1995 18:48:55 GMT ] > > > >Arno Kamphuis (kamphuis@fys.ruu.nl) wrote: > > > >Hi, > > It has been awhile since I was intouch with Linux development. > >Has Linux been now able to support multiprocessors, say a dual > >486DX4-100 Motherboard? > Let me make a guess: NO If it runs under the Intel SMP standard, then the answer is YES! I've heard people using a lot of dual pentiums, and one case each of 4 pentiums and 3 pentiums. The Linux kernel supports up to 32 processors. -- Andrew E. Mileski -- ======================================================= Dark Matter Technologies Inc. - Ottawa Ontario, Canada! mailto:dmtech@magi.com http://www.magi.com/~dmtech/ ------------------------------ From: dmtech@magi.com (Andrew Mileski) Crossposted-To: comp.os.linux.hardware,comp.os.linux.misc Subject: Re: Does Linux support (or will) RAID 5 drives and/or Hot plugged drives ? Date: 17 Dec 1995 02:07:47 GMT Pierre Blanchet (blanchet@emi.u-bordeaux.fr) wrote: > My question is in the title. > Thank you for your responses, There is a Linux driver for Raid-1 and Raid-2. *ANY* kind of disks work (supposedly). -- Andrew E. Mileski -- ======================================================= Dark Matter Technologies Inc. - Ottawa Ontario, Canada! mailto:dmtech@magi.com http://www.magi.com/~dmtech/ ------------------------------ From: dwild@starforce.com (Derek Wildstar) Subject: Re: AMD DX4-120 Date: Sat, 16 Dec 1995 11:35:20 GMT root (root) wrote: : I have update to an AMD DX4-120..noticed one problem..System won't boot a : floppy kernel unless I turn off the internal cache (bios function)...I don't : really need the function unless I have to reinstall Linux ..I had to repair an : error and needed to boot from a floppy to get into the system when I discovered : it.. : : May just be a quirk of my motherboard/bios or could be the DX4/120.. : : Anyone have similiar experience? : : The kernel starts to uncompress and then gives a crc error and halts.. : : Works fine with lilo and loadlin.. : I was able to kludge around this problem by changing my memory speed from "Fast" to "Slow" (one step down) This speed is relative to the CPU external bus speed, which you probably raised from 33 to 40mhz, so it shouldn't actually slow your system down much. I have been looking for another motherboard that will support the AMD-specific wait states and internal write-back cache, I'm not getting nearly the performance I expected. Does anyone have any suggestions, and possibly an email address/phone number/web page or the like I can consult? -- Joel Boring aka Derek Wildstar This message is not to be sent over the Microsoft Network. ========================================================== SCAN SCAN SCAN!! $10 up to 10, $.50 each thereafter high quality, 8/15/16/24-bit. email me for details! ------------------------------ From: jmcc@m5.vi.ri.cmu.edu (Jason Mcmullan) Subject: Re: 1.3.45 kernel memory usage Date: 13 Dec 1995 19:47:30 GMT Stephen Davies (scldad@sdc.com.au) wrote: : Last night I upgraded to 1.3.45 and when I booted it, I got the folowing : messages: : "Uncompressing Linux... : large kernel, low 1MB tight... : done." : How significant is this? : Can/should I do something about it? Rather insignificant. This simply means that you had a lot of drivers, IP options, whatever, compiled into your kernel. The kernel will still boot fine and run well with this message, it's just a warning that the kernel is _almost_ too large to fit into the 640k real-mode area. To remove this message, try makeing some of your drivers as modules. Modularizing the sound driver is especially helpful. -- Copyright 1995 Jason McMullan; all rights reserved; license for the Microsoft Network to distribute this text for US $100,000 per copy is hereby granted; all other use of this text by the Microsoft Network, including storing, relaying, or reuse is expressly prohibited; distribution by the Microsoft Network indicates full acceptance of the licensing terms. MICKEY# 4729-304-8732 ------------------------------ From: yku03006@yorku.ca (Blue Angel) Subject: Always IN-2000 SCSI card Date: 17 Dec 1995 02:40:28 GMT Does anyone know if Linux can handle the Always IN-2000 SCSI card? -- _ /| Allen Shymchuk a.k.a. Blue Angel \o.O' Scarborough Canada, eh! =(___)= -=[ yku03006@yorku.ca ]=- U ack! pffffft!!! -=[ alpal@yucc.yorku.ca ]=- Home of the Telefunkin U-47... York University - Computer Science ------------------------------ Crossposted-To: alt.satellite.tv.crypt From: mibu@scrum.greenie.muc.de (M.Buchenrieder) Subject: Re: Linux ISO 7816 emulator (was: New Voyager) Date: Fri, 15 Dec 1995 11:59:01 GMT koenig@tat.physik.uni-tuebingen.de (Harald Koenig) writes: >no bad idea if this really works for *ALL* CMOS clock versions and the >interrupt line isn't used for other hardware (which IRQ line is the >CMOS clock connected to ? couldn't find anything in my PC docs). Hmm - isn't that irq 8 ? Michael -- * Michael Buchenrieder * mibu@scrum.muc.de / mibu@scrum.greenie.muc.de * ************************************************************************** * Our continuing mission: To seek out knowledge of C, to explore strange * * UNIX commands and to boldly code where no one has manpage 4... * ------------------------------ From: habibie@catevr.fiu.edu () Subject: [Kernel Module] Upgrade To ELF Date: 16 Dec 1995 22:11:43 GMT Hi, I have downloaded a copy of rangelan2 module package and amd hoping to be able to use it on my linux box. According to the author of this software, the rangelan2 module source was compile under a.out and runs with no problem. I was able to compile the module with Linux kernel 1.2.13 (ELF) and installed it using insmod. Unfortunately, the module does not function with this kernel 1.2.13. If I compile the module with kernel 1.2.11 (non-ELF) and installed is using insmod, it works. I am starting to wonder if the kernel supports ELF format, does the source code of the module need some modifications to be able to run in ELF? Please kindly reply through eMail. Thank you very much. -- habibie@catevr.fiu.edu ------------------------------ ** 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-System-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux.development.system) via: Internet: Linux-Development-System@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-System Digest ******************************