Received: from SOUTH-STATION-ANNEX.MIT.EDU by po7.MIT.EDU (5.61/4.7) id AA02720; Sat, 24 Feb 96 15:35:19 EST Received: from SENATOR-BEDFELLOW.MIT.EDU by MIT.EDU with SMTP id AA08269; Sat, 24 Feb 96 15:34:57 EST Received: (from root@localhost) by senator-bedfellow.MIT.EDU (8.6.13/2.3JIK) id PAA06457 for Linux-Development-System-Dist@senator-bedfellow.mit.edu; Sat, 24 Feb 1996 15:14:56 -0500 Message-Id: <199602242014.PAA06457@senator-bedfellow.MIT.EDU> From: Digestifier To: Linux-Development-System@senator-bedfellow.MIT.EDU Reply-To: Linux-Development-System@senator-bedfellow.MIT.EDU Delivered-To: Linux-Development-System@senator-bedfellow.mit.edu Precedence: bulk Date: Sat, 24 Feb 96 15:14:50 EST Subject: Linux-Development-System Digest #420 Linux-Development-System Digest #420, Volume #2 Sat, 24 Feb 96 15:14:50 EST Contents: ncpfs probs in 1.3.68 (Andrew Ross) Re: Where is 1.3.59 (was Re: Kernel Change Summary 1.3.60) (Michael Elizabeth Chastain) Re: Massive bug in GCC 2.7.2 (David Fox) Re: need secure OS to entrust millions to (Bryce) Re: New real-time system call: nanosleep (Markus Kuhn) Re: Firewalling/Masquerade with 1.3.6x Kernels (Robert Stockmann) Re: linux/kernel/exit.c : do_exit() (Kevin L McWhirter) Re: 1.3.67 -- beginning to look a lot..like..beta (Yang-Cheng Hsiao) LATEST STABLE REV. ??? (Ken Adams) ipfw and kernels > 1.3.57? (John Grana) Re: any DHCP server port to linux ??? (Rick Hicks) Re: AIC7xxx driver problem (Greg Wood) Re: Serios problem with e2fsck (1.01), Possible EXT2FS (1.3.66) (Aurel Balmosan) ---------------------------------------------------------------------------- From: anr1001@hermes.cam.ac.uk (Andrew Ross) Subject: ncpfs probs in 1.3.68 Date: 23 Feb 1996 11:58:11 GMT Hi, I've just compiled 1.3.68. It all went ok and booted fine but I'm now having problems with the ncpfs and ncpmount. I'm using kerneld and ncpfs is a module. ncpmount bails out with a ncp_read_super: invalid wdog socket error when I try to mount a drive. The module is loaded ok - I've checked that. It appears to be a module / kerneld problem though as 1.3.68 works fine on a friends computer with ncpfs in the kernel. Has anyone else seen this or got any ideas? I'm looking into it myself but no joy yet. This was working fine with 1.3.59. Thanks for your help Andrew ============================================================================= Andrew Ross Email anr1001@cam.ac.uk Churchill College Tel (01223) 331594 Cambridge CB3 0DS http://www.chu.cam.ac.uk/home/anr1001/home.html ============================================================================= ------------------------------ From: mec@treflan.shout.net (Michael Elizabeth Chastain) Subject: Re: Where is 1.3.59 (was Re: Kernel Change Summary 1.3.60) Date: 22 Feb 1996 20:46:23 GMT In article , John Fletcher wrote: >In article <4frmnk$du@treflan.shout.net>, mec@treflan.shout.net says... >> >>Kernel Change Summary >>Linux 1.3.60 (22701 lines) >>Tue 13 Feb 1996 >> >> >Thank you for this. Is there one anywhere for 1.3.59? Yes. Read the very next line after the lines you quoted: My KCS archive: ftp://ftp.shout.net/pub/users/mec They're also on http://www.crynwr.com. Michael Chastain mec@duracef.shout.net ------------------------------ From: fox@graphics.cs.nyu.edu (David Fox) Crossposted-To: comp.os.linux.development.apps Subject: Re: Massive bug in GCC 2.7.2 Date: 24 Feb 1996 09:26:35 -0500 In article <4gjgor$b8v@bingnet1.cc.binghamton.edu> consp05@bingsuns.cc.binghamton.edu (SethMeister G.) writes: ] I beleive I have discovered a nice bug in the intel GCC 2.7.2. Try to ] compile a C++ program that uses STDIO functions, like printf, etc... The ] thing will segfault when you try to write to stdout. this is not true of ] GCC 2.7.0 and is independant of libc and libg++/libstdc++ -- Watch out ] for this one!!! It is not plausable for a bug to exist, since thousands of people use stdio with g++ 2.7.2. There must be a problem with your installation. -- David Fox http://found.cs.nyu.edu/fox xoF divaD NYU Media Research Lab baL hcraeseR aideM UYN ------------------------------ From: wilcoxb@cs.colorado.edu (Bryce) Crossposted-To: comp.os.linux.misc,comp.os.linux.networking,comp.unix.bsd.freebsd.misc,comp.unix.bsd.netbsd.misc,comp.unix.bsd.bsdi.misc Subject: Re: need secure OS to entrust millions to Date: 24 Feb 1996 16:58:49 GMT Reply-To: bryce@c2.org =====BEGIN PGP SIGNED MESSAGE===== I, Bryce , wrote: > > I'm writing documentation which advises banks on how to > setup an electronic banking software package on a > Net-connected, firewall-protected Intel box. Some of the > most important banks in the world will be reading this > documentation very soon. Iain Hibbert wrote: > > However, I can only suggest that you remove specific system > recommendations from the document and instead recommend that > they use an experienced Unix/Security admin person to choose, > install and _continually_administer_ the system if they want > to run a secure site. Such a person would know the options > and would install something that they would be able to keep > control of. I think Iain has the best answer yet! Thanks to everyone who followed-up and e-mailed me with many pertinent points to make. After a little deliberation it became very clear that the proper thing to do was to remove all security recommendations which were not specifically within the domain of our application. Security issues are too complex, and too important, to address with a mere line or two in a document which has a different purpose. Thankfully the question of a secure OS (as well as the questions of firewalling, networking, physical security, key distribution, access by insiders, etc etc etc) falls outisde of our domain. It was really interesting to read the specific issues (e.g. one guy said that "the linux TCP/IP stack has some serious flaws in it's buffer handling, and if your traffic is high, you may wish to consider that. Also the firewall code is buggy, and the support for well-hidden MD5 encrypted passwords is not there. s/key and kerberos do not integrate well with linux.".), and the larger issue of source-available OS'es versus commercially-supported OS'es. Although I'm a big fan of the free OS phenomenon, I would almost certainly have to go with an OS that was designed specifically for security, e.g. TIS, QNX, etc. Possibly the best of both worlds would be to convince some company or some collection of hackers (hint hint) to take a free-source OS and hack it into a highly secure form. Okay I sent this to all newsgroups because I started this thread and wanted to broadcast its resolution. Follow-ups oughta be directed to one specific group or another... Regards, Bryce =====BEGIN PGP SIGNATURE===== Version: 2.6.2 Comment: Auto-signed under Unix with 'BAP' Easy-PGP v1.1b1 iQCVAwUBMS9DovWZSllhfG25AQHW6QP/UoI00hwDsDyxyqkT4CSFbB1VUzyKODfR d0/LFGWdXQg3FNAwHTuDbn7ivMzfW3VWioLDgFBQeo22vLRzcyES/68tIvIM5Ehi W6ITnPo2XPJxZ2L2uhV/EAR4SNcPg6ONka9A/CbTWpxKGeT2jTYgl47YGwpsWggJ cOFpjyB8fos= =gmE2 =====END PGP SIGNATURE===== ------------------------------ From: mskuhn@unrza3.dialin.rrze.uni-erlangen.de (Markus Kuhn) Subject: Re: New real-time system call: nanosleep Date: 21 Feb 1996 19:28:44 +0100 Reply-To: mskuhn@cip.informatik.uni-erlangen.de mskuhn@unrza3.dialin.rrze.uni-erlangen.de (Markus Kuhn) writes: >SYNOPSIS > #include > int nanosleep(const struct timespec *req, struct timespec rem); ^^^ Sorry, the second parameter is also a pointer of course, i.e. the correct prototype is int nanosleep(const struct timespec *req, struct timespec *rem); Markus -- Markus Kuhn, Computer Science student -- University of Erlangen, Internet Mail: - Germany WWW Home: ------------------------------ From: stock@stm40.pstngw.tudelft.nl (Robert Stockmann) Subject: Re: Firewalling/Masquerade with 1.3.6x Kernels Date: Fri, 23 Feb 96 04:36:49 +0100 (MET) Bernd Eckenfels (ukd1@rzstud1.rz.uni-karlsruhe.de) wrote: : -----BEGIN PGP SIGNED MESSAGE----- : This is a Announce for BETA Software and Development Releases. : There are many questions for Firewalling in recent Kernels. The current : Version of ipfw will NOT support Kernels later then 1.3.60. It has to be : recompiled with 1.3.x Kernels to be used propperly. This binarie will NOT : work with 1.2.x Kernels or vice versa. : If you need a Firewalling Tool use ipfwadm instead. : ipfwadm1.2 requires Kernels 1.2.1 or later : ipfw-1.10 requires Kernels 1.1.x-1.3.60 : ipfwadm2.0beta1 runs only with 1.3.61-1.3.65 : ipfwadm2.0beta2 runs only with 1.3.66 or later. hmmm does this illustrate the desperate state of kernel, networking and runtime development for linux? Just read a desperate email of a ISP which ran Inn under linux: ext2 dies when a certain amount of inodes is used.. Also is the NFS sustained througput at the level still where it once started. Even with a P6 it only gives you 250 kbytes/sec. The guy switched over to FreeBSD and is happyer than ever with 800 to 900 kbytes/sec/ NFS speed. NFS serving should have a kernel module in linux for a long time. But with all this " make works, lets upload it...mwahh don't test it" the evil comes in deeper and deeper.. Robert : (Note: 1.3.66's masquerade support is broken, get a fix in : ftp.xos.nl/pub/linux/ipfwadm/patch-1.3.66-masq or wait for 1.3.67) : New Features: : insert or append rules, via-interface-address and name. : The new firewalling code for ipfw is currently under development. : ip_blocking split into ip_input and ip_output chains for better control. : Informations in ipfwad can be found on http://www.xos.nl/. General : Informations on the net-tools Package (including ipfw) can be found on : http://www.inka.de/sites/lina/linux/net-tools/. : Greetings : Bernd : - -- : (OO) -- Bernd_Eckenfels@Wittumstrasse13.76646Bruchsal.de -- : ( .. ) ecki@lina.{inka.de,ka.sub.org} http://home.pages.de/~eckes/ : o--o *plush* 2048/93600EFD eckes@irc +4972573817 *plush* : (O____O) If privacy is outlawed only Outlaws have privacy : -----BEGIN PGP SIGNATURE----- : Version: 2.6.2i : iQCVAwUBMSyLdIQRll5MupLRAQHrhgQAodXEs3f49W/oRF/5O17B8e2zHISriZ3P : HS+CjXPsnAa/0gcGengWYkCjGjH+TDks6A3Pf2V3WXvKZM82P2cJRFHQoyz4ThDn : r8dQ7ogoH4waCPOZx1eOZh6YxaDmtuGnNHD4zv0JIqcf2oxqH+AJDuZ67uCwkbST : oObJ2XFRG6A= : =E6ZD : -----END PGP SIGNATURE----- : -- : This article has been digitally signed by the moderator, using PGP. : Finger wirzeniu@kruuna.helsinki.fi for PGP key needed for validating signature. : Send submissions for comp.os.linux.announce to: linux-announce@news.ornl.gov : PLEASE remember a short description of the software and the LOCATION. -- ++---------------------------++----------------------------------------++ || R.M. Stockmann || Delft University of Technology || || stock@cpt7.stm.tudelft.nl || Department of Chemical Engineering || || phone: +31 15 2784395 || Section Industrial Catalysis || || home: +31 162 436177 || Julianalaan 136 || || fax: +31 15 2784452 || 2628 BL Delft The Netherlands || ++---------------------------++----------------------------------------++ ------------------------------ From: Kevin L McWhirter Subject: Re: linux/kernel/exit.c : do_exit() Date: Sat, 24 Feb 1996 11:58:38 -0800 Marc Aurele La France wrote: > > In article <312D51F9.1018E82A@crl.com>, > Kevin L McWhirter (klmcw@crl.com) wrote: > > I am writing an app that will be an X front end to playmidi. > > > I have a button on the main form that performs a play/stop operation. > > > To accomplish this I have done the following: > > 1) Parent forks child to return to form processing; > > 2) child set's p-group to it's pid; > > 3) build's command line and calls system() > > 4) calls _exit() (using xforms and don't want child to close X > > connection). > > > To perform stop action: > > 1) Parent calls kill() using the opposite of the child's pid > > this sends signal to entire p-group. > > > However: > > According to linux/kernel/exit.c : do_exit(), this creates a zombie > > process out of the child. Is this really the POSIX desired effect? > > I'd say the parent's SIGCHLD handler needs to call one of the wait* > functions, or the parent should set its SIGCHLD handler to SIG_IGN (which amounts to the same thing). > > Marc. (stuff deleted) WOW! Thanks! I thought I understood where it was neccesary to use wait() and where it was unneccesary. Guess I still don't understand ALL of what wait() does. Curious thing is that I did have a SIGCHLD handler. It changes the state of the Stop button back to Play, and sets up the next Play operation if the Repeat option is set. Anywho thx for the tip! And thx to Linus T. for the kernel source code!! ------------------------------ From: hsiao@saturn.sdsu.edu (Yang-Cheng Hsiao) Subject: Re: 1.3.67 -- beginning to look a lot..like..beta Date: 24 Feb 1996 09:18:18 GMT In article <312E2E13.53E83CC6@idir.net>, vergil wrote: >..68 is good stuff. I'm using ppp and it might be my imagination, but >everything is a lot faster... I'm not dealing with leftover connections >stuck in FIN_WAIT2 after days and days. The swapping is faster and more >intelligent (IMHO). I have a fairly vanilla system, so there might be >bugs that I never get to see, but I've been very pleased with my kernel. > > >Christopher Wall >..68 is good stuff. I'm using ppp and it might be my imagination, but >everything is a lot faster... I'm not dealing with leftover connections >stuck in FIN_WAIT2 after days and days. The swapping is faster and more >intelligent (IMHO). I have a fairly vanilla system, so there might be >bugs that I never get to see, but I've been very pleased with my kernel. > > >Christopher Wall > > Hi: I just recently upgraded my kernel from 1.2.11 to 1.3.66 and pppd from 2.1.3(?) to 2.2.0e. However, it seemed to me that it became slower than before when I ran Netscape 2.0 over PPP. I could hear my hard drive read/write like crazy. Did anyone experiece the same problem? Now, I am back to my old kernel and pppd and evrything seemed to be back to normal(btw, my system is 486DX2-66 with 8 MB RAM). So, if you know anything that can fix my problem with the new kernel and pppd, please send me an email. Thanks! Yang-Cheng ------------------------------ From: Ken Adams Subject: LATEST STABLE REV. ??? Date: 19 Feb 1996 17:31:49 GMT Sorry, I know the deal with even/odd etc... but I haven't been poking around here much lately and am in dire need of knowing what was the most stable recent version of the kernel? I've got a project that I want to release in conjunction with the 1.4 kernel, and I want to begin dev testing to make sure all my old rev. compliant code works. Specifically I rely on the mmap implementation in 1.2.1, as well as the page table directory.. etc... Also, anyone know the deal with threads? Any suggestions, pointers to material in a faq, anything would be greatly appreciated. BTW : "Butte" ... the accelerated full-screen Mesa will be begin beta in middle to late April.... anyone interested (and who owns or has access to a GLINT 300SX populated graphics board) can contact me or my web page for some information on http://fantasia.eng.clemson.edu/~adamsk (VR link, gGL) the project. For those of you who suggested fixes to my pages, thanks... I often leave off the last " at the end... so bear with me :) -- +-------------------------------------+---------------------------------------+ | Ken Adams | adamsk@cs.clemson.edu | | Virtual Reality Sys. Dev. |http://fantasia.eng.clemson.edu/~adamsk| | Clemson University Computer Science | ofc: 803-656-1682 | +-------------------------------------+---------------------------------------+ ------------------------------ Crossposted-To: comp.os.linux.networking From: jjg@pt.com (John Grana) Subject: ipfw and kernels > 1.3.57? Date: Sat, 24 Feb 1996 16:01:24 GMT I have been running a system with the 1.3.57 kernel, diald and ip masq. Recently, I compiled the 1.3.67 kernel and booted. Everythings works fine except the ipfw command that setups up the ip masquerading. I get an "ioctl - wrong argument" error. So, I tried to re-compile ipfw. It looks like most of the ipfw flags have been changed/renamed. Is there a newer version of ipfw that works with the newer kernels? Thanks. John Grana jjg@pt.com -- ___________________________________________________________________________ |John Grana, Performance Technologies Incorporated jjg@pt.com| |315 Science Parkway, Rochester, New York 14620 uupsi!ptsys1!jjg| |Phone: (716) 256-0200 Fax: (716) 256-0791 | ------------------------------ From: rhicks@mo.net (Rick Hicks) Crossposted-To: comp.os.linux.networking Subject: Re: any DHCP server port to linux ??? Date: Sat, 24 Feb 1996 17:22:14 GMT On a dark and stormy night paul@wau.mis.ah.nl (Paul Slootman) wrote: >Ah! Pray tell, how do I recognize from which net the request came from? >I'm working on the bootp-2.4.3DD-whatever server at the moment, and am >having trouble doing exactly that. >The request comes in from the client as a broadcast, and does not have >any IP address yet (that's why the client's doing the DHCP in the first >place!). I have not found any way of determining on which net the >broadcast originated (or to be more precise, on which NIC the broadcast >was received on...) I must confess to not knowing the gory details >about this type of network stuff, although I can find my way round quite >reasonably... Your router needs to be setup or upgraded to support DHCP requests properly. The router will then place the network of the request in the DHCP packet so that the server can tell where it came from. CAUTION: We could not get this to work when the request had to pass through multiple routers. It seems that the network address is set at every router hop, thus corrupting the 'real' originating address. Rick ________________________________________ Rick Hicks Systems Specialist Hussmann Corporation ------------------------------ From: Greg Wood Crossposted-To: comp.os.linux.setup,comp.os.linux.hardware,uiuc.sw.linux Subject: Re: AIC7xxx driver problem Date: Fri, 23 Feb 1996 14:55:18 -0600 Mark D. Roth wrote: > > When trying mke2fs: > > aic7xxx: Target 0 underflow - wanted (at least) (15360) got(1024) count(3). > aic7xxx: Target 0 underflow - wanted (at least) (15360) got(1024) count(3). > aic7xxx: Target 0 underflow - wanted (at least) (15360) got(1024) count(3). > SCSI disk error : host 0 channel 0 id 0 lun 0 return code = 18000000 > Current error sd08:02: sense key Medium Error > Additional sense indicates Unrecoverred read error > scsidisk I/O error: dev 08:02, sector 800 > > These error messages go on for quite a while, giving the first line > many times with various values. > > The controller is an Adaptec 2940 PCI. The machine is a Gateway 2000 > Pentium-133. I've had this problem under both Linux 1.3.46 and > 1.3.64. > > I don't think it's a hardware problem, because Win95 (which was > shipped with the machine) talks to the drive just fine. > > Is this a driver problem, or do I have something configured wrong? > Has anyone else had this problem? Any info on this is appreciated! The 2940 driver is flaky at best - are you using this on a hard drive or removable media? Target 0 underflow means that the SCSI bus is not passing the information to the device at the right speed. I would go into the bios and change the maximum transfer rate lower, and see what happens. Maybe turning off sync negotiation on the device, too. I dunno if I would do this on a hard drive tho. (this generally refers to optical disks) Just some suggestions. ------------------------------ From: aurel@uni-paderborn.de (Aurel Balmosan) Subject: Re: Serios problem with e2fsck (1.01), Possible EXT2FS (1.3.66) Date: 24 Feb 1996 16:51:39 GMT Hello, Now I can specify the problem a little better. I have started from scretch using a totally new partition table using again fdisk-3.04. The partition table is: xylo-root |2402 16:23| 211#fdisk -l /dev/hdc Disk /dev/hdc: 16 heads, 63 sectors, 2484 cylinders Units = cylinders of 516096 bytes, blocks of 1024 bytes, counting from 0 Device Boot Start End #cyls #blocks Id System /dev/hdc1 * 1 21 21 10584 4 DOS 16-bit FAT <32M /dev/hdc2 22 836 815 410760 83 Linux native /dev/hdc3 837 914 78 39312 82 Linux swap /dev/hdc4 915 2483 1569 790776 83 Linux native Every thing works fine while copying the data from the source harddisk to the new partitions. I made after every step a e2fsck. Than I copied data from /dev/hdc4 to /dev/hdc2 using: tar cf - usr | (cd /; tar xvvf -) (/dev/hdc2 was mounted on /usr) After that I unmounted /dev/hdc2 and /dev/hdc4 and rerun e2fsck with following result: ============ xylo-root |2402 16:20| 209#e2fsck -f /dev/hdc2 e2fsck 1.01, 30-Oct-95 for EXT2 FS 0.5b, 95/08/09 Pass 1: Checking inodes, blocks, and sizes Pass 2: Checking directory structure Pass 3: Checking directory connectivity Pass 4: Checking reference counts Pass 5: Checking group summary information /dev/hdc2: 15133/102816 files, 240696/410760 blocks xylo-root |2402 16:20| 210#e2fsck -f /dev/hdc4 e2fsck 1.01, 30-Oct-95 for EXT2 FS 0.5b, 95/08/09 Pass 1: Checking inodes, blocks, and sizes Duplicate blocks found... invoking duplicate block passes. Pass 1B: Rescan for duplicate/bad blocks Duplicate/bad block(s) in inode 30908: 129919 129919 Pass 1C: Scan directories for inodes with dup blocks. Pass 1D: Reconciling duplicate blocks (There are 1 inodes containing duplicate/bad blocks.) File /backup/proc/kcore (inode #30908, mod time Fri Feb 16 21:53:10 1996) has 2 duplicate blocks, shared with 0 file: Clone duplicate/bad blocks? yes Pass 2: Checking directory structure Pass 3: Checking directory connectivity Pass 4: Checking reference counts Pass 5: Checking group summary information Fix summary information? yes Block bitmap differences: -129887. FIXED Free blocks count wrong for group 3 (1553, counted=1552). FIXED Free blocks count wrong for group 15 (7, counted=8). FIXED /dev/hdc4: ***** FILE SYSTEM WAS MODIFIED ***** /dev/hdc4: 37048/197880 files, 508983/790776 blocks xylo-root |2402 16:25| mount /usr xylo-root |2402 16:25| /usr/bin/vi Can't create temp file... Does directory "?usr/tmp" exist? ============== Because the same problem occured as described in my last posting I assume that sector address calculation in the IDE-driver may have a bug. As you see in the protocol the e2fsck was ok for /dev/hdc2 (/usr) and only /dev/hdc4 was corrected. But data in /dev/hdc2 were modified as you see in the unsuccessfuly /usr/bin/vi call. Before creating the file-system I have checked them using: badblocks -svw And no bad block was detected. Also I disabled the hdparm calls in the rc-files and any swap was disabled too. I realy need some help with this. My next turn will be only creating one partition as file-system and one swap partition. If you have any suggestion plese responde to this new-group because until this problem is not fixed my system will not receive any mail or news. Thanks in advance, Aurel Balmosan -- ========================================================================== Aurel Balmosan----- aurel@xylo.owl.de ------------------------------Fnord! ========================================================================== ------------------------------ ** 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 ******************************