Re: DVD player
On Mon, Dec 03, 2001 at 09:03:26PM +0100, Christian Schoenebeck wrote: Can you also give me some hint which programs I can use for ripping DVDs or can you point me some docs on the web to get more information about that? mplayer (http://www.mplayerhq.hu) plays DVDs (needs libdvdread, and libdvdcss from videolan) and damn near anything else, if you've got the libs. On my box, it's not as good / fast as vlc (and I may have to try ogle again given others' responses). And it only sort of has a GUI. (It looks pretty, you can open a file and hit play / seek. :) However, they've just added an encoder to it, which seems to work well. Right now it's just DivX4, but that's what you probably wanted anyway. :) Eventually it'll have more feechurz; it seems to be in fairly heavy development. HTH, Mike McGuire
Re: Xine-dvdnav skipping problems
On Mon, Nov 26, 2001 at 10:41:33PM +0100, Romain Lerallut wrote: Thus spake Mart van de Wege on Mon, Nov 26, 2001 at 08:32:47PM +0100: Using the dvdnav plugin for Xine, dvd playback skips a lot. It will play a few scenes, the drive light goes on for more data, and at that moment playback halts for a couple of seconds. All it does is print the following message to the console: '406 frames delivered, 1 frames skipped, 367 frames discarded'. ... I've got the same thing, mostly with dvdnav (can stop for 30sec sometimes) and slightly with dvdnav (stop for 1/2 seconds). This is with kernel 2.4.10 and 2.4.14, however it works used to work lawlessly with kernel 2.4.3 (at least nothing noticeable). (xine* 0.9.4 on Duron 700, Via chipset, GeForce 2MX) I can possibly run a few tests if someone needs it, including reviving my old 2.4.3. Anyway, if somebody has a fix, cc: to me, please. eh. Just an thought, but since you asked: If it worked in 2.4.3 and doesn't work in the later kernel versions, it could be an issue with the VM. (heh. I don't even know quite what that means. :) I do know it deals with I/O, and from the description of it, the dvdnav plugin sounds like it does some (non-sequential) reads from the DVD to get the info for the navigation stuff. ...and all sorts of I/O related pauses and outright hangs have been blamed on the VM lately. It might be fixed soon- from my occasional look at the LKML archive just yesterday there was some discussion on it. You could try waiting a week or two and then get the latest kernel that doesn't have any known severe bugs (like fs corruption, that'd be bad...) HTH, Mike McGuire
Re: Virus incident
On Wed, Nov 21, 2001 at 06:11:22PM -0700, Josh Everist wrote: Well I can take responsibility for the messages from '[EMAIL PROTECTED]', however if the Debian User mailing list didn't send out viruses in the first place, there wouldn't be the automated reply. Then the automated reply is broken. As far as I know, this list doesn't filter anything- with the possible exception of blocking an address after repeated annoyances, such as your automated mailer. So, regarding that, it really shouldn't be sending junk mail to the receipients of the message. Replying to the sender is understandable. Beyond that, if you feel it's your duty / responsibility / whim to warn your users of a virus, then at least have it only automatically reply to *your users*. Let other people deal with the problem themselves. However, all of what I've just said would be irrelevant if this remailer was at all sane and used the actual addresses the mail was being delivered to in the first place instead of spamming anyone that happened to be in the To: field of the message. The same goes for the YODA messages. In all honesty I've merely kept up with my software and virus updates and don't actually know why my software automatically responded. If you don't know what your software is doing and why, then you aren't keeping up nearly well enough. However, if there were no response, it wouldn't be nearly as obvious that the email from [EMAIL PROTECTED] was viral. Thanks. The HTML-and-binary-attachment-only messages on a Linux mailing list weren't obvious enough. The 5 or so warnings per original helped a lot. (And if there's anyone out there these days that still needs such blatantly obvious warnings has bigger problems to worry about than email viruses.) So, if there's any way the mailing list can filter out the viruses in the first place, you won't see the automated messages. At any rate I hope people go after the real culprit(the sender of the virus) and don't take it out on us. First, multiple replies per original message is annoying. Second, others have suggested the replies are badly formatted as well, regarding standard email protocols. Finally... If A - B - C, and C needs to be prevented, stopping either A or B will solve the problem. You're apparently capable of fixing B, but apparently not willing to do anything about it. I can filter my mail. I can also complain / rant / flame. And so I have. Thanks for making your problems ours as well, Mike McGuire
Re: GTK for Mplayer?? (Possible issues w/ GTK Ximian?)
This seems to be a problem with the Ximian packages, which I can't really help you with. But I can tell you that last I checked, the mplayer GUI is rather incomplete and uh, sucky. It didn't seem to do much but play, pause, and look nice. :) Maybe they've got it doing more by now, but IMHO it's not that necessary. As for the issues with installing the gtk dev packages... Sorry, but the following packages have unmet dependencies: libgtk1.2-dbg: Depends: libgtk1.2 (= 1.2.7-1) but 1.2.10-ximian.20 is to be installed libgtk1.2-dev: Depends: libgtk1.2 (= 1.2.7-1) but 1.2.10-ximian.20 is to be installed Depends: libglib1.2-dev but it is not going to be installed E: Sorry, broken packages That's what I got.. and also I'm running potato, and ximian on my machine.. can anyone help me so I can use GUI in mplayer? ...it seems to me that if Ximian ships a special version of libgtk1.2, they should do the same with libgtk1.2-dev. Also, just try installing libgtk1.2-dev instead of gtk-1.2 or whatever it was, that and what it depends on should be all you need. Or wait for anyone who knows of any issues with Ximian on Potato with the gtk libs to speak up... HTH, Mike McGuire
Re: mp3 --- wav
On Sat, Sep 22, 2001 at 02:09:19PM +, John Griffiths wrote: found a somewhat ugly way to fake it, seems to be working though with -w the -s option talks about headerless data in the man page: The decoded audio samples are written to standard output, instead of playing them through the audio device. This option must be used if your audio hardware is not supported by mpg123. The output format is raw (headerless) linear PCM audio data, 16 bit, stereo, host byte order. forgive my ignorance, is that a .wav file? nope. WAV's have headers. So headerless wouldn't be it. Although I think a .wav is just a fairly simple header tacked on to that raw PCM data; it's possible to generate a .wav header. Also, I think some cd-burning software (quite likely if it's linux :) can burn WAV or PCM or some other format. WAV should always work, tho. Short answer: if -w works, go with it. :) HTH, Mike McGuire
Re: LS-120 grub boot with 1.44MB but not 120MB floppy: Geom Error
Yo. I've got one of these fun little things too. One of these days I'll get grub up and running, maybe, but at present I'm sticking with lilo. I've got a LS-120 boot disk to run lilo, but it needed some extra parameter. I had to add the following to lilo.conf: disk=/dev/hdb bios=0x00 I don't use grub much yet, so I don't know if there's a corresponding option for it, so here's the relevant section from man lilo.conf so you have an idea what to look for: disk=device-name Defines non-standard parameters for the specified disk. See section Disk geometry of user.tex for details. Especially useful is the `bios=' parameter. The BIOS numbers your disks 0x80, 0x81, etc. and it is impossible to decide which Linux disk corresponds to which BIOS disk (since this depends on the BIOS setup, and on the type of BIOS), so if you have an unusual setup you need to state the correspondence between Linux disks and BIOS disks. For example, disk=/dev/sda bios=0x80 disk=/dev/hda bios=0x81 would say that your SCSI disk is the first BIOS disk, and your (primary master) IDE disk is the second BIOS disk. I'm guessing the bios=0x00 tells lilo to use whatever disk the BIOS thinks is the floppy drive. But if that's what setting (fd0) was supposed to do you might be stuck. Hope you're just missing something. HTH, Mike McGuire
Re: ssh braindamage (protocols, RSA auth) FIXED
meh. keeping this short. I'm an idiot, kinda. :) SSH is a complicated mess, that's my excuse. Basically, this sums it up: RSA = v2 (*not* v1) RSA1 = v1 DSA = v2 So, when I said it was doing v1 when it should be v2, I meant it was doing RSA when it should be doing DSA. And then realized that that was the default, and it was doing what it was supposed to. eh. Reading the fine manual: HostKeyAlgorithms Specfies the protocol version 2 host key algorithms that the client wants to use in order of preference. The default for this option is: ``ssh-rsa,ssh-dss'' ...which is a *client* problem, when I'd been looking at the server all along, and not quite knowing what was going on anyway. :) Just so you guys can stop wondering what the hell I'm going on about, :) Mike McGuire
Re: ssh braindamage (protocols, RSA auth)
On Mon, Sep 17, 2001 at 10:32:36AM +0200, Pietro Cagnoni wrote: ... i've always been able to solve my ssh problems using sshd -d and ssh -v doh. yeah, I probably tried that. but now I've poked at it some more. sshd -d -d -d, all protocols enabled: not very interesting. just so you know what keys it finds: debug1: private host key: #0 type 0 RSA1 debug1: private host key: #1 type 1 RSA debug1: private host key: #2 type 2 DSA and then ssh -2 -v -v -v: eh, dumped to logs and ran diff. w/ all protocols enabled, w/ only 2 (only let it see the DSA key) 17c17 debug1: Remote protocol version 1.99, remote software version OpenSSH_2.9p2 --- debug1: Remote protocol version 2.0, remote software version OpenSSH_2.9p2 36c36 debug2: kex_parse_kexinit: ssh-rsa,ssh-dss --- debug2: kex_parse_kexinit: ssh-dss 53,54c53,54 debug1: dh_gen_key: priv key bits set: 122/256 debug1: bits set: 1016/2049 --- debug1: dh_gen_key: priv key bits set: 135/256 debug1: bits set: 995/2049 62c62 RSA key fingerprint is eb:a9:6b:36:7e:16:33:d7:38:80:48:61:c7:44:4f:e3. --- DSA key fingerprint is d0:51:d4:c5:b7:a5:93:de:05:aa:c1:ac:87:c3:a7:7a. Yes, I actually have to hide the ssh_host_rsa_key to get it to use v2. Just setting 'Protocol 2' isn't enough, it still falls back to protocol 1 and finds the RSA key. Even though the debug output shows prot v2.0. At first I thought the protocol v1.99 threw it off, but the third case with the setting 'Protocol 2' produces this output, among others: debug1: Remote protocol version 2.0, remote software version OpenSSH_2.9p2 debug2: kex_parse_kexinit: ssh-rsa,ssh-dss RSA key fingerprint is ... ...so now I think that's it. Seems to me to be preferring protocol 1, ssh-rsa, if it's available regardless of the 'Protocol' setting. Which seems to me to be broken behavior. Which I'd file a bug report for, if I wasn't worried that I'm just being an idiot. eh. Again, any comments? Anybody else trying to figure out why protocol 2 won't work? :) Mike McGuire
ssh braindamage (protocols, RSA auth)
I've been having a couple annoying problems with ssh, current version in unstable. First off is, sshd absolutely refuses to use protocol v2, either with the default 2,1 or explicitly setting 2,1 or using ssh -2 to connect. The only way I got it to use v2 was to disable completely v1, which I really shouldn't have to do. Second is something me and some friends of mine have been banging our heads about for a while; that's that RSA Authentication also just refuses to work. Make the keys with ssh-keygen, move the public key to $HOME/.ssh/authorized_keys on one of the hosts, and from another host use the private key. And nothing happens. Falls through to ask for a password. My config files are the stock ones from the package, now that I've given up on messing with them. Anybody else having problems? Or an idea? Or can confirm that it's just broke? Mike McGuire
Re: ssh braindamage (protocols, RSA auth)
Eh. Ok. With some more banging on it, RSA auth sort of works now. Before, I don't think protocol v1 was working with RSA auth, and v2 didn't work at all. However, disabling v1 altogether and generating new v2 keys for RSA auth works. Connect. Yay. Happy. But the protocol resolution or something still seems a bit broken, as this was a bit of a mess to get working. And Protocol 2,1 never would do what it was supposed to and always fell back to v1. Still interested in comments if anybody's got any. :) Mike McGuire
Re: MPEG video trouble Radeon 64 MB DDR
On Sun, Sep 09, 2001 at 01:15:13PM -0500, nick lidakis wrote: System: Tyan Tiger 230 w/dual 1 GHZ P3 512Mb ram 3com NIc 3c590 Ensoniq 1371 soundcard OS's Debian woody, debian SID, win98 SE via moblie rack HD swap Problem: Newly purchased Radeon 64MB VIVO works with openGL with all 3 OS's. While trying to play any MPEG (DVD, mpeg1, mpeg2) using any client (Videloan 0.2.03, latest xine, etc), video plays fine but when switching desktops the last frame played still remains on all my desktops... My guess, since no one else wanted to take a shot at it, is this: The DVD players probably use something called overlay or another method to get faster access to the video hardware. Specifically, (I think) overlay involves letting the app do whatever it wants to a chunk of video memory; so I'm guessing overlay support still has a few quirks to be worked out. :) If it bothers you that much, the players should have an option to turn off overlay, or whatever other acceleration method it's using. You could try playing with the settings, you've probably got a fast enough machine to do without it. HTH, Mike McGuire
Re: Compiling lame
On Mon, Sep 10, 2001 at 05:08:37PM +0200, Julio Merino wrote: Hi all, I want to compile LAME, but I can't... I've downloaded the lame 3.89 beta version (the latest) and doesn't compile. 3.88 gives me the same result, and 3.70, that is the stable one, doesn't compile too... This is the ./configure output, which breaks: eh. Why not use the debian scripts they've been nice enough to provide? :) Really, I've been using them for a while. Makes things nice and easy. Install the dpkg-dev package if you need it, check the config file debian/rules in the lame source tree, then run dpkg-buildpackage -us -uc Works great except for a few hackish things I've had to do with a few odd configure options, but I may be able to help with that if you get to it. Also, I'd reccommend sticking with the latest versions. They've been adding and tweaking things constantly, and haven't ever managed to break anything that I've noticed. :) a few random comments... Script started on Mon Sep 10 17:03:35 2001 juli:/tmp/lame-3.89# ./configure ... checking whether gcc accepts -g... yes checking dependency style of gcc... mkdir: cannot create directory `confdir': File exists ./configure: cd: confdir: No such file or directory Odd. Never seen that before. gcc checking for ld used by GCC... /usr/bin/ld checking if the linker (/usr/bin/ld) is GNU ld... yes checking for BSD-compatible nm... /usr/bin/nm -B checking whether ln -s works... yes not updating unwritable cache ./config.cache ./ltconfig: ./ltconfig: No such file or directory configure: error: libtool configure failed Stupid question: got write access to the build directory? HTH, Mike McGuire
Re: Compiling lame
On Mon, Sep 10, 2001 at 07:42:22PM +0200, Julio Merino wrote: juli:/tmp/lame-3.89# dpkg-buildpackage -us -uc dpkg-buildpackage: source package is lame dpkg-buildpackage: source version is 3.88-0 dpkg-buildpackage: source maintainer is Ingo Saitz [EMAIL PROTECTED] dpkg-buildpackage: host architecture is i386 debian/rules clean /usr/bin/dpkg-buildpackage: debian/rules: bad interpreter: Permission denied Weird. I have permissions (have checked already the directory/files) and I'm running that as root... doh. Including executable permissions? Should have mentioned that. debian/rules needs to be executable for dpkg-buildpackage to run. It's probably a feature that you have to change the perms, maybe it assumes you'll have looked at the file too? :) HTH, Mike McGuire
Re: q ad mp3-software
On Thu, Aug 30, 2001 at 12:56:54AM +0200, Robert Waldner wrote: Hi! I´m on a rather slow machine, a Pentium running at 50 MHz. This is just fine for working, and should be reasonably fast for playing mp3´s. With mpg123 (stable) CPU-usage is ~ 70 %, but I get random clicks and other nasty distortions. Ok, renice mpg123 to -3, so it should be served before anything else (other than swap et al). No change. xmms is out of the question anyway, it eats 50 % CPU on my P3/850MHz-notebook, so I don´t wonder that it doesn´t live up to the task on the slow box (not even when setuid root´ing it and using the realtime priority-thingie). A friend of mine's got a setup where he wrote a wrapper or something for mpg123 where he sets it's priority to something ridiculous. Higher than 'nice' will set it, I think he said it preempts the kernel in some places. ;) Needless to say, this makes the box useless for anything else while playing mp3s. :) His is a plain old Pentium laptop, not sure if it was that slow even (50MHz). Yeah. So you can look into scheduling system calls, sounds fun. :) HTH, Mike McGuire
Re: bad superblock recovery help?
On Tue, Aug 28, 2001 at 01:41:37PM +1000, Serge Rey wrote: ... i tried a reboot, to see if i could get into the windows partition but when i selected that stanza in lilo, lilo hangs at L? (again, this used to work prior to this recent problem). i can boot into linux just fine. That seems a bit odd... the lilo manual doesn't mention an L? error condition, and booting linux ok would imply that LILO loaded ok as well. Could be you've got something configured incorrectly or in some bizarre way. i'm thinking something got hozed on hda1 during my recent cdburning session? but i'm not sure how to proceed. i was (perhaps stupidly) using the windows partition to store the images mastered under xcdroast - this seemed to cause no problems previously, but perhaps this is a possible source of my problem. If you've got a windows bootdisk handy, and have DOS fdisk.exe, you can blow away LILO and reset to a DOS MBR. The command to do this is: FDISK /MBR . (That's assuming there isn't a backup from before using LILO.) Just make sure you still have a way to boot linux, like another bootfloppy. Well, I'd suggest taking a good look at the LILO docs no matter what you do, and hope nothing got really b0rken. (And if all else fails, reinstalling windows is SOP anyway. ;) HTH, Mike McGuire
Re: The Sound of Silence
On Mon, Aug 27, 2001 at 09:33:21AM +0900, Curt Howland wrote: Sorry for the multiple replies, but I have some new data. Adding to the audio group didn't work, because the SB modules never loaded. I have figured out finally that the sound card modle is ES1869, but I have not yet found any information on manufacturer other than this model number. If anyone has information on this model, I'd apreciate it. Now, off to do a Google search... I almost suggested you use the maestro driver, but that isn't it. (ESS1968 is maestro, ES1869 is something else. lovely, ain't it?) But anyway, if you haven't found it yet, the ES1869 seems to be supported by the sb_ess driver. This is in the kernel, ver 2.4.9, but I think it's been there a while if you're still doing 2.2.x. Digging some more: sb_ess isn't a driver itself (in 2.4.9 at least), it's part of the sb driver (100% SB Compatibles), which I think puts you back where you started. :) There's some extra documentation for the ESS chips in {kernel source tree}/Documentation/sound/, which more or less says everything should be configured automatically on inserting the module, but if it doesn't there's other stuff to try. HTH, Mike McGuire
Re: AW: ipmasq support in potato kernel
Well, I called my original post random speculation, so maybe I should just shut the hell up. :) But not before one last shot at this... On Thu, Aug 23, 2001 at 10:23:30PM -0500, Nathan E Norman wrote: On Thu, Aug 23, 2001 at 11:11:09PM -0400, Mike McGuire wrote: On Thu, Aug 23, 2001 at 09:31:59PM -0500, Nathan E Norman wrote: [this was getting too long. and it's not terribly intersting it is. if you've got to know, you should be able to find it.] b) IF you have the special superFOO deluxe masquerading module in the kernel THEN you don't need ipmasq, and You're looking at it backwards. If you install ipmasq (the package), your kernel must have firewalling support compiled in. Otherwise the ipmasq package is useless. Of course, you don't _need_ to install ipmasq to use the functionality you've compiled into the kernel; it just makes it easier. OK, I know that ipmasq needs kernel firewalling support. However, I don't think I made clear there's also kernel masquerading support *in addition* to the firewalling support... So you've got firewalling and you want masquerading. There's a few choices: 0) kernel firewalling, write your own masq scripts w/ ipfoo. 1) kernel firewalling, use ipmasq's scripts. 2) kernel firewalling and masquerading, turn it on w/ ipfoo. Just so you don't yell at me again ;) here's how #2 works. (2.4.x kernel using netfilter/iptables) (somewhere in the kernel config) CONFIG_IP_NF_IPTABLES=m # general filtering CONFIG_IP_NF_FILTER=m # packet filtering CONFIG_IP_NF_NAT=m # Netwk Addr Translation CONFIG_IP_NF_NAT_NEEDED=y CONFIG_IP_NF_TARGET_MASQUERADE=m# special NAT module Then to set it all up, one whole line (though this just does the masquerading, no real firewalling, but I think that the masq stuff was the point to this whole mess...) : iptables -t nat -A POSTROUTING -o ppp0 -j MASQUERADE As I understood it, the default kernel just does filtering, and maybe NAT. And that's all ipmasq needs to work. Perhaps the masquerading module is just some extra feature in 2.4.x/iptables? (That could explain most of this confusion. arg. (But then, me being completely wrong could too. I hope that's not it. ;) ) [more blather] I don't understand what you're on about with this faster stuff. IP Masq support in the kernel is ip masq support. It doesn't work automatically; you have to configure it. The ipmasq package does exactly that. eh. When you said you'd recommend a custom kernel especially if you're manipulating packets, I just assumed you thought it was faster too. And I still think the difference is that it's *not* IP masq support in the kernel when using ipmasq; it's IP masq in user space using the kernel IP filter support. The IP_FOO_MASQ module I mentioned above is obviously kernel IP masq support, so if I'm right about all that- and I hope I am- kernel code should be faster vs. user space code. But then, I could be wrong. Quite likely, in fact. If I am, just tell everyone that I'm an idiot, and why, and as simply as can be done to prevent eating away any more of their bandwidth. Now I've got tv to watch, and I've got to throw some stuff together so I can leave for school tomorrow, so I'm done. No more. :) Respectfully submitted, Yours etc., and this time I quit. Really. :) Mike McGuire
Re: ADSL problem with large packets
On Thu, Aug 23, 2001 at 12:59:38PM +1000, Andrew Pollock wrote: Hi, I just got ADSL running on my home network with a Debian gateway (two ethernet cards). I think I've got an MTU problem. Large transfers just don't work, but I can make an SSH connection out (for example) and pinging and the likes works. I've got pty pppoe -I eth0 -T 80 -m 1412 and mtu 1412 in my peers file for the DSL connection as per documentation. I haven't messed with the MTUs on anything behind the Linux box. Should I? I thought the whole idea of this MTU clamping was this wasn't required? eh. I think I remember that the default MTU is 1500, so the machines behind the gateway would get their packets fragmented, which can lead to (bizarre, seemingly random, and hard to diagnose) problems. I'd suggest you try setting everything to 1412 anyway, just to check. HTH, Mike McGuire
Re: Help - Large Files Support
On Thu, Aug 23, 2001 at 04:30:32PM -0500, Indraneel Majumdar wrote: On Thu, Aug 23, 2001 at 09:25:00PM +0200, Federico 'Derfel' Stella wrote: Check headers in /usr/include/linux/. If UTS_RELEASE in version.h is 2.4.x you have nothing to do. I am running unstable (upgraded from potato via woody) on Linux-2.2.5 version.h: #define UTS_RELEASE 2.4.8 #define LINUX_VERSION_CODE 132104 #define KERNEL_VERSION(a,b,c) (((a) 16) + ((b) 8) + (c)) cannot generate file 2GB I remember someone saying that both the kernel and the program need large file support. I *don't* remember if you said which program and version of said program you're trying (and probably the libs on which it depends, too), so if you did, sorry. Try asking again, anyway, with that info. Sorry I can't help you further, but I don't a thing about it other than that. :) HTH, Mike McGuire
Re: Upgrade
On Fri, Aug 24, 2001 at 09:20:38AM +1000, Joel Mayes wrote: GECOS == GECOS [EMAIL PROTECTED] writes: GECOS Is there any way to upgrade to a higher version of debian ? GECOS I always get lost in `dselect' or `apt-get' . Greg Try ``apt-get update apt-get dist-upgrade'' er... try that after you look at '/etc/apt/sources.list'. I assume you're running 'potato', as that's what now is 'stable' and is what gets installed by the bootdisks. So, you're probably either thinking of: A) When does Debian v2.x or v3.0 or something come out? If this is your question, check that the lines in sources.list say 'stable' and not 'potato', so that when 'woody' becomes 'stable', the entire system should be updated rather painlessly, though requiring a rather large download of new packages. And this should be happening Real Soon Now. ;) or, 2) All my software is out of date. I want new stuff. In this case, you can try 'testing' if you're feeling brave, or 'unstable' if you're feeling brave, lucky, etc. :) If you just want newer stuff, you could put 'woody' in sources.list, and it will eventually become stable. If you want / need the latest stuff, that would be unstable. I think it's called that to scare people away even though it is fairly stable, since occasionly something goes TERRIBLY HORRIBLY WRONG. ;) If you decide to go to testing or unstable, you might want to wait a week as there's been some glitches in the upgrade of some important packages that aren't quite fixed yet. HTH, though I think I got carried away and this is probably all explained on www.debian.org somewhere. :) Mike McGuire
Re: AW: ipmasq support in potato kernel
On Thu, Aug 23, 2001 at 04:07:34PM +0200, Peter Palmreuther wrote: Hello Andreas, On Thursday, August 23, 2001 at 2:37:11 PM, you wrote (at least in part): or does the standard kernel support masquerading? I've installed a Debian 2.2 (w/o 'r') shipped with kernel 2.2.17, am using this kernel and have installed a masquerading firewall, so the standard kernel coming with Debian should support it. I'm using package 'ipmasq' for starting / stopping firewall and I've not seen it loads a separated module for activating this functionality. eh. This is just a guess, but I'm guessing that the ipmasq program uses the standard firewall rules to do masquerading. Or you can do a kernel compile and enable the module / option for masquerading in the kernel. There shouldn't be any real difference between them, other than with the kernel support a simple rule would do the job without needing the ipmasq package, and might be slightly faster. HT(random speculation)H, :) Mike McGuire
Large Files (was Re: something or other)
eh. Maybe I should set mutt to set the Reply-to or whatever to the list... whoever that was may want to send that again to the list as I deleted it already. anyway, they mentioned using a C test program on assumedly a large file, and apparently fopen was failing with an error about large files not being supported. Checked some things; nothing on the fopen(3) manpage about large files, but open(2) has this (not sure it's relevant, but something to consider): O_LARGEFILE On 32-bit systems that support the Large Files System, allow files whose sizes cannot be represented in 31 bits to be opened. OK. fopen() is the library call, and open() is the system (kernel) call, and I'm guessing the one calls the other. However, I have *no idea* how exactly any of this information relates to something practical. Any gurus out there? :) I do have a hunch tho. It could just be a compiler option or something that's needed. HTH, Mike McGuire
Re: AW: ipmasq support in potato kernel
On Thu, Aug 23, 2001 at 09:31:59PM -0500, Nathan E Norman wrote: [my garbage snipped] Bzzzt. ipmasq (the package) is simply a collection of SCRIPTS that, depending on which kernel you have installed, enables IP Masquerading via ipchains, iptables, or ipfwadm (whatever the 2.0 stuff was called). The ipmasq package does _not_ include any modules or other code which actually does the masquerading (since htat's the job of the kernel. eh. That's what I meant, I think. :) Start again: ipmasq uses the standard kernel support of ip(fwadm|chains|tables) to do the same thing as the special kernel masquerading module for ip$1. That's a lot clearer, and more concise, than the mess I originally posted, at least if you can understand the pseudo-regexps. :) The conclusions still stand, though: 1) the result is the same b) IF you have the special superFOO deluxe masquerading module in the kernel THEN you don't need ipmasq, and iii) the module should be faster. or it might play Yankee Doodle through the PC speaker when someone uses SSH. IIRC, a potato install leaves you with a kernel that does have ipchains support; however, I always recommend compiling a custom kernel, especially if you're manipulating packets. Same here. If it's faster, and I expect it would be, that's better. Of course, if you're just masqing one box over a modem, I doubt it would make a difference. But compiling kernels is fun. :) TMTOWTDI, YMMV, HTH, HAND, etc. :) Mike McGuire
Re: Encrypted Filesystems
On Wed, Aug 22, 2001 at 02:05:41PM +1000, Damon Muller wrote: Quoth Rog?rio Brito, I'm also interested in encrypted filesystems and I think that the only way to go (reasonable one, that is) is to use the international kernel patch available at http://www.kerneli.org/. I remember seeing a project recently which was a kernel module which allowed you to have encrypted loopback devices with patching the kernel of using the international kernel patch. Not sure if this is it, but I've been playing around just with the international patch using a loopback device (in the standard kernel). (I compiled them all as modules, but I think you meant some other non-standard kernel patch / module.) Anyway, there's a howto on linuxdoc.org, Loopback-Encrypted-Filesystem-HOWTO. HTH, Mike McGuire
Re: Encrypted Filesystems
On Wed, Aug 22, 2001 at 10:13:02AM +, Iain Smith wrote: Anyone happen to know the current state of the crypto patches? I haven't seen a new patch for 2.4 since april... for 2.4.3! Yeah, that's it, I guess. Maybe they got it right the first time. :) One minor problem (too minor to bother fixing, apparently) is that the kernel makefile changed slightly somewhere between 2.4.3 and now. So the patch fails on a chunk of Makefile. Editing by hand works. If you're trying and don't quite know how to do that, here's what it should look like: ... CORE_FILES =kernel/kernel.o mm/mm.o fs/fs.o ipc/ipc.o NETWORKS=net/network.o CRYPTO =crypto/crypto.o LIBS=$(TOPDIR)/lib/lib.a SUBDIRS =kernel drivers mm fs net ipc lib crypto DRIVERS-n := ... HTH, Mike McGuire
Re: how to make exim run queue only when ppp link is up?
On Wed, Aug 22, 2001 at 02:21:19AM +0100, Stig Brautaset wrote: Diald keeps bringing up the link on request from exim. I have only had this problem since I made a fresh install about a week ago. Earlier, the exim cron-entry that run the queue was controlled by a set of scripts in diald's ip-up.d and ip-down.d directories, but not so now. Why is that? I could of course move the thing myself, but I want to do this the debian way. I have tried dpkg-reconfigure exim, but that didn't work... I think what you're looking for is the exim cron job that runs every 15 minutes. I'd look into disabling that, but I don't remember if there's some special Debian way to edit cron entries. Here's where it is: '/etc/cron.d/exim'. I *think* if you change it, it either won't be updated or it'll ask you to when exim inevitably gets updated. I probably would have looked into it when I was messing with diald a while back, but I just gave up instead. Not to discourage you or anything, I'm just lazy. :) HTH, Mike McGuire
Re: Why so big(2)
On Tue, Aug 21, 2001 at 03:51:16PM +0100, P Kirk wrote: /dev/hda2 703M 574M 93M 86% / du -m /usr/399 du -m /usr/src/ 173 du -m /usr/share148 du -m /home 120 I don't have anything other than samba and python installed. How can /usr/share be so big? ... tetex-base install tetex-bin install tetex-lib install This stuff is pretty huge, and you probably don't need it. (Anybody know a reason? And one good enough to have it installed by default, if that's the case?) Anyway, on my box: spoon:~# du -s /usr/share/* | sort -nr | head -n 10 -- (top 10!) 66956 /usr/share/texmf-- (created by tetex on install) 48076 /usr/share/doc 17064 /usr/share/locale 12184 /usr/share/man 10316 /usr/share/lyx -- (my reason for keeping tetex) 7732/usr/share/terminfo 6876/usr/share/i18n 6616/usr/share/perl 5128/usr/share/games 5028/usr/share/zoneinfo ...and tetex-base itself is about 30MB. So you'd probably free another 100MB if you can't find a reason to keep it. The other big stuff in /usr/share is more or less required- doc, locale, man, terminfo, perl, and games of course, for fortune. ;) HTH, Mike McGuire
Re: sshd
On Tue, Aug 21, 2001 at 03:13:30PM -0400, Titus Barik wrote: chronos:~# ipchains -L ipchains: Incompatible with this kernel I am running the 2.4.6 kernel (ipchains only works with the 2.2 series?). Not exactly, but it's not supported by your kernel. So it's not that. ssh -v -l barik 128.61.40.17 returns: OpenSSH_2.5.2p2, SSH protocols 1.5/2.0, OpenSSL 0x0090601f debug1: Seeding random number generator debug1: Rhosts Authentication disabled, originating port will not be trusted. debug1: ssh_connect: getuid 0 geteuid 0 anon 1 debug1: Connecting to 128.61.40.17 [128.61.40.17] port 22. debug1: connect: Invalid argument Eeeww. Could be you missed something configuring your kernel, but I can't think what off the top of my head. You can try running strace on the ssh connection; should be as simple as : strace ssh -v -l ... and seeing what it's doing that gets the Invalid argument error. HTH, Mike McGuire
Re: IDE errors on mounting DVDs
I'd appreciate any comments on the following problem I'm having when I mount a DVD: VFS: Disk change detected on device ide1(22,0) UDF-fs DEBUG lowlevel.c:57:udf_get_last_session: XA disk: no, vol_desc_start=0 UDF-fs DEBUG super.c:1410:udf_read_super: Multi-session=0 hdc: command error: status=0x51 { DriveReady SeekComplete Error } hdc: command error: error=0x50 end_request: I/O error, dev 16:00 (hdc), sector 1799392 This is on 2.4.8 x86, on a Dell laptop with a TORiSAN DVD. The DVD appears to be mounted, and the volume descriptor is read successfully. I can move around the filesystem and read many files, but any attempt to read a VOB file produces the IDE errors again, and the reading process gets an I/O error. Initially I thought this might be a region coding issue, but my drive is set to region 1 and this is a region 1 disc. Close there, you forgot the other annoying thing with DVDs - the so-called encryption. I'm almost certain that's what produces the errors; I get it on my box when I try doing things like that. Or when trying to play a DVD in a player that I didn't / couldn't / forgot to compile a CSS decoder for. You can't read an encrypted DVD by itself since DeCSS isn't in the kernel. (Yet? ;) Most / all Linux DVD players handle it themselves, or with a library, though it seems to me each project has its own library. You'll probably have to compile the player yourself as Debian doesn't package the decryption code, even in non-US (arg). It'll be a lot of fun. Yes. Anyway, a few players that I know of: OMS: http://www.linuxvideo.org/ VLC: http://www.videolan.org/ Xine: eh, search for it on freshmeat. maybe XMPS also. VLC and Xine are in Debian, minus the CSS, but it's easy enough to do apt-get source and compile a CSS-enabled version. OMS has a deb line for apt/sources.list on their site. Take your pick, try 'em all. Find one that works best. HTH, Mike McGuire
Re: ipmasq/ipchains/reading TFM
On Mon, Aug 13, 2001 at 01:05:25PM -0400, Stephen Nosal wrote: Folks - I'm running a potato box (2.2.19) as a firewall/gateway for my tiny (2 machine) home network. Most everything is working fine, but my internal network box can't see certain web pages that are visible from the gateway ... 1) Am I going about this the right way - does it sound like ipmasq? 2) Can someone point me to the right section of the docs on this, I'm wading through pages of ipmasq docs but can't seem to put my finger on this. I had what I think is the same problem, was driving me nuts. From the ipmasq-HOWTO I found at http://ipmasq.cjb.net/ 7.15. ( MTU ) - IP MASQ seems to be working fine but some sites don't work. This usually happens with WWW and FTP. . A perfectly good way to bypass this is to change your Internet link's MTU to 1500. Now some users will balk at this because it can hurt some latency specific programs like TELNET and games but the impact is only slight. On the other hand, most HTTP and FTP traffic will SPEED UP! (And then explains how to change it for Linux, Windows, etc.) HTH, Mike McGuire
Re: IP masquerade not working!! :( (fwd)
On Sun, Aug 12, 2001 at 08:23:40PM -0400, dude wrote: On Mon, 13 Aug 2001, Roberto Diaz wrote: Does any of this work directly from the gateway box without any MASQ rules loaded? Thank you for your help.. it is some kind of problem with my network cards... It is quite strange.. if I try to send a huge file (using ftp) from my host to the gateway it goes at the speed of light.. but if I try to take a file from the gateway to the host it crawls.. :??? This is not ip-masquerade specific.. but somebody has some idea? Sounds like your gateway does not know its hosts That could be it. Seems to me though that it would only cause it to be slow to connect, but be fine afterwards? I'm not sure on this, but it would help to know what part is slow. If it connects fine and is just slow, it could be any number of other things. Check your logs. Always. :) And look at the settings of course; you may be right that it's the cards. Just a shot in the dark- you mentioned in another message the mtu was set ok, just check that it's the same on both boxes. Sorry I can't be of more help. Maybe if you gave some info on the cards somebody else could help a bit more. HTH, Mike McGuire
Re: Searching in dselect
The '/' key is helpful in searching for a string in the package name, but it will not search the descriptions of the packages. I seem to remember in a previous version (even before slink) that it used to do this. Hm. Not quite what you were looking for, but 'apt-cache search' does this. As long as apt knows about packages, and if dselect uses apt as its get method or whatever, it does. As for searching descriptions in dselect, I'd like to know too. I'm surprised one of the dselect evangelists hasn't piped up on this yet... ;) HTH, Mike McGuire
Re: HELP broken dpkg ( previously HELP - installing broken libglib1.2-dev)
[cruft deleted] Ouch. I've never had to deal with a system as loused up as that one. But perhaps it can be solved with a: dpkg --force-all -P libglib-dev (libglib1.2-dev libgtk1.2-dev libgtk-dev) Hopefully this will clean out the system making way for a fresh install of everything for glib1.2 and gtk1.2 Let the list know if this still doesn't work. Maybe someone who has solved this problem will also come across the thread. --mike On 10 Aug 2001 11:31:35 -0700, Jatin Golani wrote: Hi Michael, Thanks for responding. To answer your questions: a) Yup I'm root. b) a dpkg --force-remove-reinstreq --purge libglib1.2-dev, has the same fate :( I don't know how to see which directories it's trying to write toif I could then I could manually do itdo you know how I can do that?? Also even if I delete the files manually how do I then tell dpkg that the package has been deletedeven after a purge it keeps showing that the package is installed though broken. Any help would be appreciatedI can't understand what's happening...would appreciate help. Bye --- Michael Heldebrant [more cruft] I've got an idea, but it's a bit ugly and probably should only be used as a last resort. (Unfortunately you appear to be getting to that point...) First, do : dpkg -L pkgname and remove (by hand, or with a script if you're feeling a bit brave :) the files it lists, for each of the broken packages. Then : echo pkgname purge | dpkg --set-selections to get it out of dpkg's database. Of course, you might want to wait a bit so anybody else can figure out something better, or say why this is a bad idea... Anyway, I *think* that much of dpkg will work, even with the broken packages. HTH, Mike McGuire
Re: CrushLink Welcome!
On Wed, Aug 08, 2001 at 08:43:59PM -0600, [EMAIL PROTECTED] wrote: OK. Who signed in? Are you going to post the password? Probably some sick twisted individual, like the one that signed up the LUG list at my college... as someone else said, could be Bill Gates. ;) But anyway... NNN Yeah, this thing's amusing / annoying. The admins of the LUG list tried to get removed, I think, but every once in a while we get some confused email from someone (who also probably didn't sign up either) who's been told CWRULUG has a crush on them. Argh. Mike McGuire
Re: libmp3lame.so -- where?
On Wed, Aug 08, 2001 at 11:07:19PM +0100, James Green wrote: I've been battling to make 'vcr' work for some time now, and it's moaning about a lack of libmp3lame.so at runtime. The library is allegedly found in lame, which isn't in Debian. I downloaded the source tarball, created a deb from it, only to find it only creates libmp3lame.a, not .so. I looked at converting, and it looks fairly horrible. Hm. I *think* there's a configure option to build it... but looking through the Makefile it doesn't seem to get built by just 'make'. You could try building it again if you don't still have it around, and then try 'make libmp3lame/libmp3lame.so', that's in there. The library is apparently also in avifile, which is broken: Sorry, but the following packages have unmet dependencies: avifile: Depends: libqt2.2-mt (= 2:2.2.3-3) but it is not installable E: Sorry, broken packages Eh. You running testing? I don't think it's in stable, and in sid it needs libqt3. And it sounds like something that would happen with testing. :) You could use dpkg-deb and just extract libmp3lame.so from it tho. HTH, Mike McGuire
Re: CrushLink New Password
On Thu, Aug 09, 2001 at 04:35:08PM +0100, P Kirk wrote: Does debian-user have an admin who junks this kind of stuff or do I need to write a procmail rule? Yeah, since there doesn't seem to be a way to get the hell off the damn thing when you log in. arg. Should probably look for an abuse addy for this crap, and/or there probably is an admin for the lists. But I'm busy. Somebody else do it. ;) Mike McGuire
Re: libmp3lame.so -- where?
Hm. I *think* there's a configure option to build it... but looking through the Makefile it doesn't seem to get built by just 'make'. The configure option --enable-shared is on by default, and thus probably is broken someplace along the way. Yup. Noticed that. Found this in a makefile... You could try building it again if you don't still have it around, and then try 'make libmp3lame/libmp3lame.so', that's in there. [make errors] Aha. Just tested this. It's in Makefile.unix, so you need to tell make that: make -f Makefile.unix libmp3lame/libmp3lame.so This builds with lame from cvs ~10 minutes ago, v3.90a6. Don't know if it *works* or not. :) The library is apparently also in avifile, which is broken: Sorry, but the following packages have unmet dependencies: avifile: Depends: libqt2.2-mt (= 2:2.2.3-3) but it is not installable E: Sorry, broken packages ... Hm. Actually it's avifile-player that needs libqt3. My dpkg doesn't know anything about just 'avifile'. But both(?) libavifile and libavifile-dev have a libmp3lamebin_audioenc.so[something]; just making a symlink might work. And downloading the package and getting the lib out with dpkg-deb shouldn't be a problem either, other than having things floating around that dpkg doesn't know about. :) HTH, Mike McGuire
Re: port forwarding
On Sun, Aug 05, 2001 at 02:36:45PM -0400, Michael P. Soulier wrote: Hey people. I'm trying to set up port forwarding to permit file sharing with napster from behind my firewall. So, looking up a friendly howto, I then entered this: rabbit:~# ipmasqadm portfw -a -P tcp -L 24.114.252.76 6699 -R 192.168.0.2 6699 portfw: setsockopt failed: Invalid argument In my experience, the cryptic Invalid argument error means there's something missing in the kernel that the command requires. I don't know what ipmasqadm needs; if it runs on top of ipchains I think you need at least a 2.2.x kernel. (2.0.x has ipfwadm, 2.2.x has ipchains, and now 2.4.x has iptables (with backward compatibility for the other two)). You might want to read some docs, this kind of thing should be explained. HTH, Mike McGuire
Re: proftpd : 2 anonymous areas?
On Mon, Aug 06, 2001 at 03:27:24PM +0100, Rory Campbell-Lange wrote: I'm trying to make 2 or more anonymous areas on our intranet where people can only view data, not write it. The two directories are /home/x/y/z and /home/x/a My three strategies: 1) make an anonymous directive with suitable directory directives within. Problem: everyone can see /home/x If the permissions on the directories are set up correctly, this shouldn't be the case. Ex.: drwx-x /home/x(chmod 701) drwx---r-x /home/x/a (chmod 705) ...will allow access to /home/x/a without allowing access to /home/x for all users. So, for some other user, 'ls /home/x' will result in Permission denied, but 'ls /home/x/a' works; the user just has to know that /home/x/a exists. This is generally how it's done, I think. At my university, everyone in certain CS classes has an account, and has to have a web visible directory (like a in your example) without giving everyone access to their projects. I don't really know much about the other stuff you posted; I'd suggest trying the permissions first. HTH, Mike McGuire
Re: FW: Careful. This is for information only.
| Amazing! Someone on /. proposed writing a script that whenever | anyone hits your web server with , you automatically connect | back and halt the attacking machine, thus stopping the spread. I would ike to see something like this grin. I was thinking of putting a CGI script as /default.ida on my apache server and doing some funny stuff with it -- maybe combine the previous auto-alert stuff to notify the system it is hosed. Now if windoze had 'mail' and an MTA this shell access could allow one to mail the admin (as root of course) right after the probe. Does anyone know how to kill a process (ie the worm or IIS) or shutdown a windows system from the command line? Ooh! Ooh! And I never thought this would be of any use whatsoever. :) Ok, in Win95 at least, this works (pretty sure this is right): rundll32 user32.dll,ExitWindows w00t. Mike McGuire
Re: FW: Careful. This is for information only.
rundll32 user32.dll,ExitWindows Meh. Doesn't quite work, even on 95. For one, it's ExitWindowsEx. And then it only shuts down if you click OK. And then it only really logs out if there's more than one account. There might be some other way like that that works, but it's probably a lot easier to just make something crash. ;) Mike McGuire
Re: DNS [subtopic: Re: Internet Connection Sharing, was: Re: Ethernet]
On Fri, Aug 03, 2001 at 12:06:46PM +0100, Ade Talabi wrote: Alvin Oga, How do you turn on ip forwarding on machin A, which is a WIN 98SE box I *think* Win98 can actually do that (95 can't, IIRC), but I don't know how. I'd ask why you'd want to do that instead of using the Linux box to do the forwarding. :) (It was easy enough for me, but someone in another thread had the obvious and annoying problem of Win-only drivers. If that's the case, well... sorry, man. ;) On a related note, according to the subject at least, is there a way to do a minimal / fake DNS without installing BIND? ie: gets a DNS request, if it doesn't know it passes it on to a real (ISP's) DNS and caches the result. hm. I think I remember hearing the term caching DNS, is that what I'm thinking of? Mike McGuire
sites not responding
First off, it's *NOT* that ECN thing, I didn't put it in the kernel. And it's not a DNS issue, or the site's problem (well, it *could* be) because (arg) it works in windows. And it shouldn't be my iptables setup, since the only thing there is a single masquerade rule (I haven't really started to mess with this yet. :) Anyway- www.pbs.org won't connect. Neither will www.newscientist.com. There's probably others, these are the only ones I've noticed, or the ones I've noticed since I realized they weren't just down. Traceroute on www.pbs.org gets to what seems to be their gateway, and dies. I'm at a loss. Can anybody think of some other reason this is happening? Mike McGuire
Re: sites not responding
Ack. Re: Karsten's mail, I get the same IP from host, and the trace stops at that same pbs-gw one. So I try a browser again, which... works. I was certain I'd tried that, but I guess not. So the problem now is that the *masqed* machine can't get to www.pbs.org. Looks like I'll be messing with iptables some more sooner rather than later... Mike McGuire
Re: sites not responding - FIXED
Ack. The problem was mismatched MTU/MRU settings between the masqed and masqing boxes. Ugly, nasty, stupid stuff. It's in the IP Masq FAQ I've been using; seems to be some hosts don't like responding to fragmented packets. The writers of the FAQ imply the blame lies with these sites: ...because they are filtering ALL FORMS of ICMP (including Type4 - Fragmentation Needed) messages in a fray of security paranoia, they are breaking the fundamental aspects of the TCP/IP protocol. Oh well. Things work now. Yay. And I got to have a whole thread talking to myself. Woo. :) Mike McGuire
Re: using convert to batch convert
On Wed, Aug 01, 2001 at 01:14:06PM +0100, Andrew Pritchard wrote: Quoting Christopher S. Swingley [EMAIL PROTECTED]: The command I would expect to work: convert -sample 50% * Another method that I often use is: $ find ./ -name '*.gif' -exec convert -geometry 50% {} {}.png \; The {}'s are replaced by the filenames that 'find' found, one at a time. So this command would reduce the size of all GIF images in the current (and all sub-) directories, as well as convert them to PNG files. Correction to the above: Don't use -geometry to change the scale, use -sample as orginal poster said! find ./ -name '*.gif' -exec convert -sample 50% {} {}.png \; Otherwise it complains that it can't find the files! Took me a little while to work out why it was overwriting my files with a 0 byte file :) That teach me to RTFM - thank god for backups. Yeah, that always sucks. But anyway, doesn't that end up with files named like 'foo.gif.png'? (Assuming the guy even wanted to convert to .png in the first place. hm. :) Fun with bash: # for i in *.gif; do convert -sample 50% $i ${i%.gif}.png ...or even this, if he wants to keep gifs: ${i%.gif}_s.gif The former changes foo.gif to foo.png, the latter to foo_s.gif. Yay. May need quotes if there's spaces, may need some work in ridiculous cases, etc. It took me forever to find this, and I was actually looking for it because there were things that DOS could do (like ren *.doc *.txt) that I couldn't figure out how to do in Linux / bash. :) I think I found it in a script since man bash is a little incomprehensible. Take a look if you're brave. :) Mike McGuire
Re: Adaptec Raid
So most of the patch was applied except for Hunk #2 which failed at 735. the file Makefile.rej looks like the following. [EMAIL PROTECTED]:~/src/linux$ more drivers/scsi/Makefile.rej *** *** 727,730 sd_mod.o: sd.o sd_ioctl.o $(LD) $(LD_RFLAG) -r -o $@ sd.o sd_ioctl.o --- 735,741 sd_mod.o: sd.o sd_ioctl.o $(LD) $(LD_RFLAG) -r -o $@ sd.o sd_ioctl.o + + dpt_i2o.o: dpti.c + $(CC) $(CFLAGS) -c dpti.c -o dpt_i2o.o How can I get these lines in the correct place in the Makefile? should I insert them manually at the line positions 727-730 and 735-741? Would this be a problem with the version of kernel(2.2.14) that I am patching? Hrm. What happened was patch tried to add the lines starting with +, but apparently didn't find the lines above to know where to put it. You can look at the file (drivers/scsi/Makefile) for lines similar to those, and put the added lines following those. Probably, there's been some small change and the lines don't appear exactly as the patch expects them to. HTH, Mike McGuire
Re: Adaptec Raid
Now when I run the patch command: [EMAIL PROTECTED]:~/src/linux$ patch -p0 ../patches/dpt_i2o-2.0-2.2.18 I recieve the following message can't find file to patch at input line 4 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -- |diff -urN linux-2.2.18/drivers/scsi/Config.in linux-2.2.18.SuSE/drivers/scsi/Config.in |--- linux-2.2.18/drivers/scsi/Config.inMon Dec 11 01:49:42 2000 |+++ linux-2.2.18.SuSE/drivers/scsi/Config.in Mon Feb 26 17:46:21 2001 -- File to patch: And it is prompting me for a file to patch. do you know how to specify the kernel to patch or even what the name is and where it is located in the src. I think you just need to change the -p0 option to -p1, which strips off the first part of the path (the linux-2.2.18[.SuSE] part) when applying the patch, and if you're in the linux directory everything should work since it will look for drivers/scsi/Config.in, etc. HTH, Mike McGuire
Re: lilo does not work!
On Mon, Jul 30, 2001 at 05:18:58PM -0300, Miguel Griffa wrote: Hi from the manual: LI The first stage boot loader was able to load the second stage boot loader, but has failed to execute it. This can either be caused by a geometry mismatch or by moving /boot/boot.b without running the map installer. could it be possible that you miss a step in compile/install kern? At 10:00 p.m. 30/07/01 +0200, Markus Hansen wrote: hi guys i tried to install a 2.2.19 kernel inseead of a 2.2.17. after booting it sayed LI LI LI etc. Seems to me this is a fairly widespread problem, as I'm getting it too, but with kernel 2.4.7. I've read through the bugreport (here: http://bugs.debian.org/105788 ) and someone thought that there was an issue with the newer kernels handling the disk geometry in some new way that messed things up (the former problem in the manual in the previous message). However, I've tried booting an older (2.2.17) kernel and running lilo under that with no success, so I personally think there's a problem with lilo, which I can't check because there isn't an older version still on debian's ftp servers. I've also tried playing with the BIOS settings, lilo options like lba32, but that hasn't done anything. Right now I'm booting the new kernel off a floppy (which isn't too painfully slow with a superdisk drive, heh) Loading the kernel off the floppy and mounting the root fs from the hard drive works fine. So I tried just using the floppy to point to a kernel on the HD, but that didn't work. So, as I see it, those of us with the problem are stuck booting from floppies unless they've got some other working install, and waiting for someone who knows better to figure out what's going on. Joy. :-/ Mike McGuire
Re: lilo does not work!
On Mon, Jul 30, 2001 at 08:07:02PM -0400, Mike McGuire wrote: ...stuff... So, as I see it, those of us with the problem are stuck booting from floppies unless they've got some other working install, and waiting for someone who knows better to figure out what's going on. Joy. :-/ Mike McGuire Ignore that guy, he doesn't know what he's talking about. ;) Anyway, a bit more fiddling and it just seems to work now. In the BIOS, I'm not sure that I actually changed anything, but now it's either set to use LBA for the drives, or automatically using LBA anyway. And I've got lba32 in lilo.conf. And I don't think it works without it. So, as I see it, lilo's handling of LBA drives/geometry got a little more strict in the latest version. Maybe. I'm no expert at low-level HDD functionality. Hopefully, though, somebody will make enough sense out of this rambling to put it to good use. :) (In short: try turning on LBA in lilo.conf the BIOS. yeah.) Mike McGuire
Re: Arrg! I want DVD
On Thu, Jul 12, 2001 at 12:12:59AM -0600, Cameron Matheson wrote: Hey, I just compiled the latest version of oms, but when I try to play a dvd, it gives me these errors: I've a few things to suggest you check from the error messages... Using MMX for IDCT transform Using 3DNOW for motion compensation LOG_INFOplugin_mgr.c:pluginLoad#484 loading plugin: (CodecSpu) spu , (null) LOG_INFOplugin_mgr.c:plugin_find#376plugin /usr/local/lib/oms/plugins/dec_spu.so matched spu , (null) : spu , (null) LOG_ERROR input_udf.c:_dvd_read#261 read (Input/output error) LOG_INFOinput_udf.c:_dvd_read#262 most probably the disc is encrypted 1. make sure some CSS decryption was compiled / installed. I think OMS moved their CSS support to a library called libcss. LOG_INFOinput_udf.c:_dvd_read#263 please install DVD ioctls 2. check your kernel. I don't remember when this stuff was added, if you have an older 2.2 kernel you might need to upgrade / patch. That, or some other kernel-level stuff. UDF fs, maybe. LOG_ERROR nav_dvd.c:_dvd_pre_read#824 cell_pos:0 unable to get full block (-1) LOG_INFOthread_input.c:_oms_thread_input#27 no more data left on input I noticed that my kernel output is kind of weird too: end_request: I/O error, dev 16:00 (hdc), sector 1228008 hdc: command error: status=0x51 { DriveReady SeekComplete Error } hdc: command error: error=0x50 3. if nothing else worked, I seem to recall seeing something about these messages being caused by a buggy DMA driver. You can get hdparm to turn off DMA for the drive, or maybe set an option to LILO when your computer starts. In any case, check some docs. I think there's a sort-of out-of-date DVD-HOWTO out there which should still be useful for making sure you have all the necessary kernel support. HTH, Mike McGuire
Re: Disk related system hangs
(snip description of HD troubles) Up til now sounds like it could be a flaky driver, or HD, or chipset... I have not had this problem while under other OS's on the same machine. ...but if that's the case it's probably a driver. I'm running woody with kernel 2.4.2 built from the kernel source package. The disks are IDE; my main working disk is Maxtor13G, with 33 as its peak xfer rate (this is from my memory). I have a Gigabyte GA 71XE4 motherboard, with AMD chips (756 is probably the relevant one). Athlon 800 MHz CPU. Is this any kind of a known kernel problem? Or does it ring any bells? The symptoms make me suspect this is a software problem, though perhaps that's being optimistic. Since you've got an AMD, you might / probably have a VIA chipset, which I've heard problems about. I don't have one, so I can't say for sure, and you should probably search for more info on it, but here's what I can think of: - if you can get into linux, and have hdparm installed, you can try turning off DMA and stuff. this might even be possible with kernel options at the LILO prompt. - find out what IDE chipset you have, and look for it in the kernel config, and search too. http://www.google.com/linux is good. Thanks. sure thing. HTH. Mike McGuire
Re: [OT] Port numbers
...And also I'd like to know how to see which ports my computer has open. Others have suggested netstat, but nmap is another tool that can do the same thing, and has the added bonus of seeing which ports other people's computers have open. :) Mike McGuire
Re: Debian/Woody and kernel 2.4.6
On Mon, Jul 09, 2001 at 05:31:12PM -0700, Vibol Hou wrote: You'll need these tools: http://people.debian.org/~bunk/ -Vibol -Original Message- From: Central Park [mailto:[EMAIL PROTECTED] Sent: Monday, July 09, 2001 4:36 PM To: debian-user@lists.debian.org Subject: Debian/Woody and kernel 2.4.6 Hi, Just a quick question, will Debian/Woody (testing) need any extra updates to compile and use a linux 2.4.6 kernel? I want to use iptables and stuff. Email me directly, as I am not subscribed to the list. Thanks in advance Michael No, that url is for potato. Woody does not need any special tweaking. There should be no reason for woody to ship with anything other than kernel 2.4.x when it's finally released (except, maybe, for kernel 2.6 coming out by then ;^). noah Just the same, you could check Documentation/Changes in the source tree, which lists what versions of some tools are required for the that kernel. Specifically, what gcc/make is needed to compile, what versions of filesystem utilites and PPP-type stuff. Testing should have all these by now, probably. Or you can always try unstable. :) Mike McGuire
Re: Funny Story
Did you know that there are Nuclear Battlecruisers (i'm not a soldier, don't know the exact name for this type of ship) running on windows NT ? Scary huh ! heh. I wonder if that's the same kind that had a divide-by-zero error... A divide-by-zero error is hardly an operating system problem, is it ?? Is Debian possibly immune to this ?? ;-) well, no. in theory, at least. something like that. it's more a problem of bad programming / software engineering- the incident makes a damn good case for value checking and testing. :) I won't say anything about, oh... a certain operating system encouraging horribly, shoddily written, lazy, poor excuses for programming. I think that other flame war is enough for the list already (and they might have covered this already). hehehe ;-) Mike McGuire
Re: Funny Story
Did you know that there are Nuclear Battlecruisers (i'm not a soldier, don't know the exact name for this type of ship) running on windows NT ? Scary huh ! heh. I wonder if that's the same kind that had a divide-by-zero error in some navigation system or something. true story. crippled the ship and it had to be towed back to port and the software replaced, some other stuff repaired. I heard about it in this book a friend of mine was reading, something like a biography of the number zero- yeah, I know a bunch of math geeks. :) Mike McGuire
Re: Saving some real radio file
I want to save some real files, but this option is not allowed when you run real player, is there any way of doing it? I figure dumping or piping it, what would be the way of calling the executable for real player from command line? Not for streaming realaudio files (*.ra). Most *.ram files *can* be saved. There was a utility out a few years back which was supposed to allow saving a RealAudio stream to disk. It did this, IIRC, by running interference between rplayer and your audio device. Unfortunately, I could only ever save about ten seconds of audio before it would lock up or quit. hm. not sure if this is the same thing, but there is / was something called streambox vcr. unfortunately, it's windows-only, and isn't being maintained (AFAIK) and is hard to find because, well, Real sued them. heh. you might have as much luck complaining to whoever runs the site to use a different (more open) codec. ;) Mike McGuire
Re: could not autodetect X server: *discover* not found
Does anybody know where is discover or did anybody experience a similar error message in the latest xfree-xserver-update in unstable? Apparently, it autodetects the video driver X should use, and apt looks for it in the configure part of installing the X packages. You shouldn't have to worry about it since you're upgrading and it should be configured already. It's been working fine for me. HTH, Mike McGuire
anybody else having loopback problems?
yay unstable. a few days ago, I started having all kinds of problems with networking- fetchmail wouldn't fetch mail, samba not working, etc.- that I eventually tracked down to the loopback device not getting brought up on boot. did ifconfig and route for lo, and everything's fine for now. (have to put it in some init script...) the only explanation I can come up with is that some package got updated and broke something since I didn't do it. ;) but I'm not sure, so I thought I'd ask if anybody else had/has the same problem. Mike McGuire
Re: anybody else having loopback problems?
got a couple replies, good thing I already fixed it cuz they weren't much help. both said net-tools broke, and one said to upgrade to the version I already have installed. thanks anyway, and if anybody else gets random net problems and ya can't ping localhost, check ifconfig and route for lo. (yay unstable. :) next project is to get mutt so as to have people reply to the list. this one probably won't take three days to fix, gods be praised. :) Mike McGuire
Re: Cat-ting binary files to the console
Every once in a while I slip up at cat a binary file to the console. (Or just forget to give mkisofs the -o flag.) This causes the console to use WEIRD characters, just plain gibberish. Is there any way to get rid of this without rebooting? Thanks! Ben Pharr Try: # setterm -reset (and backspace over any crap it put on the command line first) And if all else fails, you've (probably) got five more terminals. :) HTH, Mike McGuire
Re: Console question (2x)
1. I have heard that its possible to run console in multiple resolutions. How do I do this in debian, specifically? 2. How can I have more console windows than the default 6? I want to bump up tty7 and above used for X. Also, what happens if I want more terminals than 12? There are only 12 function keys! I am running debian testing (maybe it matters). I am untrained in the mystic ways of the console. Any help would be appreciated. :) Put vga=ask in your /etc/lilo.conf run /sbin/lilo and you will be given a choice of resolution on boot. I added vga=ask and I got no such choice option at boot time. yes. I did run lilo after i edited the conf file. Also, I tried vga=extended and that doesnt work either. There are no errors and the 'normal' mode is always used. Here is the response from dmesg: Console: colour VGA+ 80x25 Any ideas? I'm fairly sure you need kernel support for the additional vga modes. and I think it needs to be compiled in, since this stuff happens way before the modules. anyway, from the kernel config help: Video mode selection support CONFIG_VIDEO_SELECT This enables support for text mode selection on kernel startup. If you want to take advantage of some high-resolution text mode your card's BIOS offers, but the traditional Linux utilities like SVGATextMode don't, you can say Y here and set the mode using the vga= option from your boot loader (lilo or loadlin) or set vga=ask which brings up a video mode menu on kernel startup. (Try man bootparam or see the documentation of your boot loader about how to pass options to the kernel.) HTH spoon