Subject: Linux-Development Digest #574 From: Digestifier To: Linux-Development@senator-bedfellow.MIT.EDU Reply-To: Linux-Development@senator-bedfellow.MIT.EDU Date: Wed, 23 Mar 94 14:13:09 EST Linux-Development Digest #574, Volume #1 Wed, 23 Mar 94 14:13:09 EST Contents: crash hosed my superblock...can it be saved? (Jerod Tufte) Re: 'format' program (Sun style) (Keith Medcalf) Re: IPX compliancy? (Alan Cox) PPP defroute problem - hosts, rc.inet1 config? (Eric Kimminau) Kernel don't recognizes HP PC LAN Ethernet Card ! (Thomas Heiling) Hanging CSLIP with ftp (Nick Vargish) Re: Real-Time Linux and a/d device drivers (Matthew Donadio) Re: in libc-4.5.21? (Dwayne Springfield) plotter works with 0.99pl1[4,5], but not 1.0 (Denis Endisch) bug? or don't I get something? (Filip M Gieszczykiewicz) ---------------------------------------------------------------------------- From: jet5@pyrite.SOM.CWRU.Edu (Jerod Tufte) Subject: crash hosed my superblock...can it be saved? Date: 23 Mar 1994 10:40:15 GMT I had my first crash in many weeks a few hours ago, and I can't seem to be able to do anything to recover. I was shutting down X, when the system just suddenly seized. I rebooted to get this message: extra data not valid Current error sd801 sense Key Illegal request Additional sense indicates Logical block address out of range and when I booted with an old recovery disk, I got: in2000: 01 80 4b 36 00 Dumping sense buffer: 112 0 5 0 0 0 0 10 0 0 host 0 id 0 lun 0 return code 28000000 Sense class 7 sense error 0 extended sense 5 scsidisk I/O error dev 0801, sector 2 EXT2-fs unable to read superblock. My system is a dx2/66 VESA m/b w/ IN2000 scsi and S3-805VLB video. someone please tell me there's a tool that can help me recover some/all of my files. thanks. Jerod Tufte jet5@po.cwru.edu -- jet5@po.cwru.edu Case Western Reserve University Computer Engineering jet@b62528.student.cwru.edu <<< finger me for PGP2.3a public key ------------------------------ From: kmedcalf@zealot.uucp (Keith Medcalf) Subject: Re: 'format' program (Sun style) Date: 23 Mar 1994 06:40:23 GMT Reply-To: kmedcalf%zealot@io.org W. Tait Cyrus (cyrus@jemez.eece.unm.edu) wrote: : Last week I purchased a new SCSI disk and after only 6 days it quit : working. Fortunately I had made a backup off all my stuff the day : before it died so I didn't loose a lot of data. If the SCSI disk is properly designed, and properly implements the SCSI spec, all that is required is to set the appropriate control pages, issue a mode select, the issue a format unit with the certification bit set (which is SUPPOSED to be the default). Many non-SCSI compliant drive manufacturers (particularly of low-cost low-reliability low-stability drives) do not implement the SCSI spec correctly (most notable is Maxtor). If you set the control mode pages, do a mode select and a default format unit, and the drive develops an error within 2 years, you should send it back to the manufacturer for replacement as it has a manufacturing or design defect. Manufacturers who create high quality drives will take them back for free re-conditioning under warranty on a yearly basis (CDC, Fujitsu), and will last for years of continuous heavy service with no problems. Other drives (not to mention Maxtor again, but they make the most unreliable and feature-free drives you can waste your money on) will last barely any time at all before developing errors. This is my opinion and experience only, and your mileage may, of course, vary. You will find however that a bottom-of-the-line Fujitsu drive will operate and last much longer than a top-of-the-line Maxtor, and that the Maxtor does not implement most of the features in SCSI that are most useful. I have a CDC (real CDC, not ScuzGate) drive that is now coming up on 8 years of continuous heavy service, almost 24 hours a day, 366 days a year, with nary a hitch. I have yet to see any maxtor drive last more than a year under the same load ... -- Keith Medcalf | Internet: kmedcalf@zealot.uucp kmedcalf@io.org Keith Medcalf and Associates | UUCP: ... uunet.ca!io.org!zealot!kmedcalf Database Systems Engineer | Anonymous UUCP: (416) 465-4370 in: nuucp Toronto/Canada (416)465-9578 | Fidonet: 1:250/750 - IBMNet: 40:6482/303 ------------------------------ From: iiitac@uk.ac.swan.pyr (Alan Cox) Subject: Re: IPX compliancy? Date: Tue, 22 Mar 1994 14:55:03 GMT In article <1994Mar14.185508.46244@ucl.ac.uk> zceed04@ucl.ac.uk (Mr Ivan Alastair Beveridge) writes: >Basically, I was wondering if anyone has made Linux compliant with Netware >at all. As I do not really know much about protocols, I cannot really ask >much more than this. There is a beta test IPX layer for Linux, but no netware support. Novell guards its netware details with lawyers and complex licensing agreements involving thousands of dollars. So forget it - Linux does Lan manager and NFS ------------------------------ From: ekimmina@pms709.pms.ford.com (Eric Kimminau) Subject: PPP defroute problem - hosts, rc.inet1 config? Date: 22 Mar 1994 15:42:00 GMT I have just completed upgrading to LINUX 1.0, GCC 2.5.8 & LIBC4.5.21. PPP was working beautifully on 99.14 before the upgrade. I have a static hostname and IP address on every connect, so I want to hardcode my system files to use it. I am still unable to get a stable chat script to work consistantly, but manually connecting via kermit will work with an additional step. Here lies my problem: Now, after connectiong to my ppp server and escaping back to the kermit prompt and running pppd, I now have to manually add the default gw route to get things working. Under 99.14, that route was added automagically. I have tried with and without the defaultroute command. Could someone who has ppp working in that way under 1.0 send me a couple things: #1 A copy of your /etc/hosts with the significant portions (routeer, net, localhost, etc. #2 A copy of your rc.inet1 or wherever you happen to configure lo and/or ppp0 with your hostname, ip, default raoute, net and gateway. #3 A copy of your chat (or kermit) startup script if you have it working. Thanks in advance! -- Eric Kimminau Workstation Systems Department 313-322-3431 Product & Manufacturing Systems ekimmina@pms709.pms.ford.com Ford Motor Co. Planning and Implementation "Not an official Ford Spokesperson" TIP#111 ------------------------------ From: tom@wpzd07.pzlc.uni-wuerzburg.de (Thomas Heiling) Subject: Kernel don't recognizes HP PC LAN Ethernet Card ! Date: 22 Mar 1994 15:57:30 GMT Hello all! Finally i got a new machine and tried to install Linux. Everything fine, some fiddling with XFree 2.1, and no the Problem: I have a HP PC LAN Adapter / 16 TL+ (aka HP 27252A ) and compiled the Kernel ( Version 1.0 + patch 1 + 2 ) with CONFIG HP PCLAN. But after compiling, the Kernel don't recognizes my Card. Here is my Machine: HP Vectra 486/66 MHZ, 8 MB, integrated VGA Card Localbus with 512 KB HP 27252A Ethernet Card /* This Card is mentioned in the Ethernet-HOWTO as one of the best Cards, but there are no hints / tips , why the Kernel don't recognizes it */ What i have done this weekend : HP has a nice program for setting up the card, it is called hplanset. I have done this : Set the IO to 300-31fh IRQ to 10 Station Adress is 080009-530D4C compiled with CONFIG HP PCLAN -> no chance compiled with CONFIG_LANCE -> no chance compiled with every available Ethernet card -> no chance Tried it with LILO : ether=10,0x300,0,0,eth0 -> no chance Copied a kernel from this machine ( This is a 386/25 MHz HP Vectra with only 4 MB and a HP LAN Card, the older model with only 8 bit ) to the new machine, but it fails again. And now i have absolutly no idea, what to do next. Any help would be very nice. Thanks Thomas -- =================================================== Thomas Heiling Pharmacist & Doctorate at Pharmazeutisches Institut Uni Wuerzburg - Germany Email phar006@wrzx12.rz.uni-wuerzburg.de (HP-UX) tom@wpzd07.pzlc.uni-wuerzburg.de (Linux) or phar006@vax.rz.uni-wuerzburg.de ( VAX ) =================================================== -- =================================================== Thomas Heiling Pharmacist & Doctorate at Pharmazeutisches Institut Uni Wuerzburg - Germany Email phar006@wrzx12.rz.uni-wuerzburg.de (HP-UX) tom@wpzd07.pzlc.uni-wuerzburg.de (Linux) or phar006@vax.rz.uni-wuerzburg.de ( VAX ) =================================================== ------------------------------ From: vargish@yx.sura.net (Nick Vargish) Crossposted-To: comp.os.linux.misc Subject: Hanging CSLIP with ftp Date: 23 Mar 1994 14:38:41 GMT Folks, Here's the scoop... All I have to do to hang my CSLIP connection is try to ftp an uncompressed xpilot map across the connection. Once I compress the map (which I will include here), the transfer goes smoothly. I suspect it has something to do with long strings of identical characters and the compression algorithm(s), but I'm basically taking an educated guess (no facts or code-grazing to back me up). I thought someone might be interested in this, assuming it's not some boneheaded mistake on my part. And by the way, I meant _hung_, so don't try this unless you're willing to restart your CSLIP connection... Nick Vargish p.s. This is the map I was talking about, gzipped and uuencoded. It should work gzipped, and fail ungzipped... ============================================ CUT HERE begin 644 13.map.gz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end ============================================ CUT HERE -- ---------------------- ---------------------- ----------------------- |. Nick Vargish .| |. SURAnet .| |. O: (301) 982-4600 .| |. vargish@sura.net .| |. systems engineer .| |. rust never sleeps .| ---------------------- ---------------------- ----------------------- ------------------------------ From: donadio@mxd120.rh.psu.edu (Matthew Donadio) Subject: Re: Real-Time Linux and a/d device drivers Date: 22 Mar 1994 16:42:57 GMT Kevin Brown (kevin@frobozz.sccsi.com) wrote: : There may be something I'm overlooking, of course (which wouldn't be : surprising), but I'm somewhat convinced that between the above and some : kind of semaphore implementation, you should be able to do pretty much : everything you'd expect out of a "real" threads implementation. You can emulate a lot with fork(), but in my experience the applications I have written using pthreads() seem to behave better. Mutexes can be done with semaphores, but conditions (as in pthread_condition) get a little hairy when you have a master process controlling a bunch of slave processes. Plus, pthreads lets you give up the processor with pthread_yeild() which can make threaded applicaitons more effieient. -- Beaker aka Matt Donadio | Life is short, --- __ o __~o __ o donadio@mxd120.rh.psu.edu | ride like ---- _`\<, _`\<, _`\<, --- Penn State Cycling ---| the wind. --- ( )/( ) ( )/( ) ( )/( ) ------------------------------ From: dwaynes@netcom.com (Dwayne Springfield) Subject: Re: in libc-4.5.21? Date: 20 Mar 94 18:54:06 GMT Joseph Toman (toman@darkwing.uoregon.edu) wrote: : Hi, I am trying to compile various source code packages for "Lee-noocks" :) : and I don't seem to have the ANSI C standard include file . It is : neither in Slackware 1.1.2 nor in libc-4.5.21 on tsx-11. Where can I find it? : Thanks, Johannes sometimes becuase of the gcc installation that file gets moved to a relatively obscure spot in the directory hierarchy.... if you have not tried using the find command to locate the file (find / -name stddef.h -print) then you just might find the file in a directory path resembling but not necessarily exactly like ---> /usr/lib/gcc-lib/i486-linux/2.5.7/include/stddef.h Like I said... obscure. You will of course adjust OS and version numbers according to your own installation :) hope this helps. l8r, dwayne -- +------------------------------+-------------------------------------------+ | Dwayne Springfield | Time flies like an arrow... | | dwaynes@netcom.com | Fruit flies like a banana. | | dwayne.springfield@octel.com | Groucho (Karlo's Brother) Marx | +------------------------------+-------------------------------------------+ ------------------------------ From: dendisch@cat..physics.uwo.ca (Denis Endisch) Subject: plotter works with 0.99pl1[4,5], but not 1.0 Date: 22 Mar 94 21:07:49 GMT Hi there I have a plotter (HP ColorPro) attached to /dev/cua1. It worked fine with plain 0.99pl14 and 0.99pl15, but with 1.0 it seem to loose characters from time to time. Sometimes the plots are ok, but most of the time the plotter starts going crazy in the middle of the plot, obviosly getting some wrong vector values. For the older kernels (pl14,pl15) I have NOT used the kernel patch mentioned in the Printing-HOWTO (Dec 1 93, included at the end of this posting) for setting up a serial printer at that time, since it worked fine without it. In the current Printing-HOWTO (Feb 22 94) the patch is not mentioned any more. Also the relevant lines in /usr/src/linux/kernel/chr_drv/tty_ioctl.c have changed, so that the older patch can not be applied anyway. It looks as if some kind of bug was implemented by this recent changes. Does anybody has similar problems or a idea what's going wrong? Or a idea for solving the problem? My current solution is not to use 1.0 :-( Thanks Denis ==================================================================== Some more information: I have the plotter running with lpd connected to a dummy device, as it is described in the Printing-HOWTO. The setting for the serial port are made at boottime with 'stty'. I have set up these settings with the 0.99pl14 kernel and have nothing changed since then. ==================================================================== My stty settings are: cat:~$ stty -a < /dev/cua1 speed 9600 baud; rows 0; columns 0; line = 0; intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = ; eol2 = ; start = ^Q; stop = ^S; susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; flush = ^O; min = 1; time = 0; -parenb -parodd cs8 hupcl -cstopb cread -clocal crtscts -ignbrk -brkint ignpar -parmrk -inpck -istrip -inlcr -igncr -icrnl ixon -ixoff -iuclc -ixany -imaxbel -opost -olcuc -ocrnl -onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab3 bs0 vt0 ff0 -isig -icanon -iexten -echo -echoe -echok -echonl -noflsh -xcase -tostop -echoprt -echoctl -echoke ==================================================================== From the older Printing-HOWTO (1 Dec 93) about setting up a serial printer: ... First, there is a kernel patch you need to make. This applies to pre-net-2 lpd as well (actually, the use of this should probably just be commented out of the lpd source code). The change is for at least patchlevel 10, but probably all others as well. in /usr/src/linux/kernel/chr_drv/tty_ioctl.c, around line 390, you will see: case TIOCEXCL: return -EINVAL; /* not implemented */ Change this to return 0, recompile the kernel, and you'll be all set. ... ==================================================================== I can give more info about system configuration, if this is important. -- Denis Endisch Phone: (519) 661 - 2111x6413 Department of Physics Fax: (519) 661 - 2033 Univ. of Western Ontario Email: dendisch@uwo.ca London, Ontario, Canada N6A 3K7 ------------------------------ From: filip@alpha.smi.med.pitt.edu (Filip M Gieszczykiewicz) Subject: bug? or don't I get something? Date: 23 Mar 94 16:45:10 GMT Greetings. I ran a butchered version of iozone 1.6 (old) with mods to allow for concurrent execution (to test just to what extent IDE sucks in multi-tasking applications - oh, btw, it sucks the chrome off the bumpers... more later) Anyway, for this little test to be "fair", I had to shoot Linux buffers in the foot (by allocating most of my memory with an "eatmem" program). I left about 1.5MB for the multiple instances (up to 20) of the 23K executable. What I noticed was that linux was groping the swap even though there was ~1MB in the buffers. After a few [dozen] trials, it was REALLY eating swap like there was no tomorrow! I'd like to know something before I go any further: What does linux prefer to do a) write/read directly to disk or b) write/ read to buffers which are really swap [sigh]. Also, can it tell the difference? Further, is there anything that can be done to help this? Also, my system is as follows: Linux 1.0 kernel (debugging on, SCSI compiled in (not used)), xiafs FS on a 200MB Maxtor IDE (17 ms), 486DX 2/66 VLB, 16MB ram, 13MB swap (partition), 256K cache, 11MHz ISA buss, 33MHz VLBuss. (note that eatmem program "ate" 11MB) I include a list of "free" after it started looking fishy: =================chop=with=axe================chop=with=axe===================== total used free shared buffers Mem: 13748 13668 80 1268 1536 Swap: 13368 3096 10272 total used free shared buffers Mem: 13748 13432 316 976 1864 Swap: 13368 3508 9860 total used free shared buffers Mem: 13748 13668 80 1336 2212 Swap: 13368 3768 9600 total used free shared buffers Mem: 13748 13668 80 1624 2480 Swap: 13368 4488 8880 total used free shared buffers Mem: 13748 13164 584 912 2832 Swap: 13368 4608 8760 total used free shared buffers Mem: 13748 13668 80 1408 3048 Swap: 13368 4728 8640 total used free shared buffers Mem: 13748 13400 348 1000 3180 Swap: 13368 4856 8512 total used free shared buffers Mem: 13748 13384 364 1000 3684 Swap: 13368 5376 7992 total used free shared buffers Mem: 13748 13384 364 1000 3752 Swap: 13368 5444 7924 total used free shared buffers Mem: 13748 13624 124 1008 4012 Swap: 13368 5464 7904 total used free shared buffers Mem: 13748 13584 164 836 5420 Swap: 13368 6640 6728 total used free shared buffers Mem: 13748 13668 80 1320 4852 Swap: 13368 6712 6656 =================chop=with=axe================chop=with=axe===================== By this time ~5 hours & maybe 30-40 trials had been run. Right now, as I write this, "free" says: total used free shared buffers Mem: 13748 13012 736 1056 4860 Swap: 13368 6740 6628 And, let me exit the eatmem, total used free shared buffers Mem: 13748 7268 6480 1044 4864 Swap: 13368 1120 12248 Well, the swap looks better but... why was I loosing memory above... Is that a "fact of life", is it a bug, is it a leak, is it paranoia, what is it? Also, why does linux keep buffers @ 4MB even though it [should] know that they are really swap. Or is that the problem... it doesn't know the buffers are swap... eh? BTW, I case anyone is wondering what kind of #'s I got from my benchmarks... If you're planning to get an IDE, you DEFINITELY want to read this (hopefully you'll change your mind): Well, the best (or worst, depending on your point of view) test was 20 instances of iozone writing/reading a 1MB file. RESULTS: 20 : concurrent jobs 1 : MB (file size) 10425 : bytes/second READ 89767 : bytes/second WRITE [sniff..sniff..Whhhaaaaaa....] Right behind (or in front :-) was 10 instances of 2MB iozone. RESULTS: 10 : concurrent jobs 2 : MB (file size) 20274 : bytes/second READ 76145 : bytes/second WRITE It got better with less instances... 1 instance of 20MB was RESULTS: 1 : concurrent jobs 20 : MB (file size) 302009 : bytes/second READ 552318 : bytes/second WRITE which just goes to prove that while IDE is "fast" in a single task environment like MS-DOG or, to a lesser extent, MS-WINDOZE, it really reeks for anything that has 2 or more processes writing to the disk. And, God forbid, you have 20 users on line... This is my BEFORE segment... I'll rerun all these tests after I get my BT445S & 1.8GB :-) and post the AFTER... I can upload the kludged iozone to sunsite so others can have a go at it (if you don't believe me). A more complete version of this article will be posted to comp.periphs.scsi to stop silly flame-fest on IDEvsSCSI speed... Take care (sorry this is so long and I hope the first item can solved. Thanks!) -- +-->Filip "I'll buy a vowel" Gieszczykiewicz | E-mail: filip@alpha.med.pitt.edu | ftp to bode.ee.ualberta.ca and get OLD electonics faq from /pub/cookbook/faq | Checkout: ftp bode.ee.ualberta.ca and get file /pub/cookbook/contents. Neat. | Making money with CS and spending it on EE, robotics, windsurfing, & dreams. ------------------------------ ** 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 ******************************