--- Day changed

--- Log opened Tue Jun 10 00:00:02 2003
00:06 < david> caker: around?
00:06 < caker> david: hey
00:06 < david> caker: still want me to take a look at your switch?
00:07 < david> caker: got side tracked doing other stuff Friday:-/
00:07 < caker> david: sure, but just know that I've decided to live with the ping solution
00:07 < caker> david: no worries :-)
00:07 < david> caker: that's a bodge, not a solution :-)
00:07 < david> caker: there is some reason why it's not working right :-)
00:08 < caker> david: pretty sure it points to my switch, but I would assume by default everything should work
00:08 < david> caker: assume = silly :-0
00:08 < david> caker: don't assume anything with Cisco stuff
00:08 < david> caker: my ISDN router was, by default, configured to have the bri0 device down - Useful
00:09 < caker> david: you sound confident - let's go for it!
00:09 < david> caker: can you mail me access info (PGP encrypted, if needed) so I can look tomorrow
00:12 < caker> david: ok, sent
00:13 < david> caker: thanks :-)
00:13 < david> caker: I won't break it
00:13 < caker> david: thank you - yeah not any worse than it is, i'm sure
00:15 < david> caker: so, what exactly happens?
00:15 < caker> david: any way you can activate my uml.org account, the email link didn't work (when I corrected it manually)
00:15 < caker> david: you aren't going to see it, because of the ping cron job
00:15 < caker> but, once the MAC addresses dissapear off "show mac<tab>" table, the IPs don't function
00:15 < caker> arps are going all over the place, as I traced it from everywhere (even the switch)
00:16 < david> caker: what's your username?
00:16 < caker> david: caker
00:16 < david> doesn't find you
00:16 < david> let me do it manually
00:16 < caker> prob because it was > 24 hrs 
00:17 < david> caker: hrm, ok
00:18 < caker> arp table didn't matter - it could be empty and the IPs would still work, it was the switch's mac table that made it break - which makes sense
00:19 < caker> it almost seems like the host bridge needed to be "refreshed" of what MACs it had
00:19 < david> right, since the switch is just bridging and has no concept of arping
00:19 < caker> because pinging all the IPs, not even from localhost, seemed to make it update the switch's MAC table .. ?
00:19 < caker> go figure
00:19 < david> caker: the switch should learn MAC addresses based on traffic, but maybe the packets are being dropped somewhere
00:20 < caker> Let me rephrase that - pinging the IPs *while they still worked* kept the mac tables working
00:21 < caker> if they timed out, not even pinging from the host would bring it back
00:21 < caker> i had to log into the uml's console and generate traffic
00:21 < david> caker: strange
00:22 < david> caker: config looks fine on the switch, so I'll have a think tonight
00:22 < caker> david: ok .. thanks
00:22 < david> caker: no idea wtf is up with uml.org - I'll look at that too
00:22 < caker> david: heh
00:40 < Pahan>  00:39:19 up 22 days, 11:06,  7 users,  load average: 3.30, 2.06, 1.12
00:40 < Pahan> heh heh
00:40 < Pahan> Two UML machines apt-get upgrading
00:40 < david> Pahan: using skas or tt?
00:40 < david> that's nuts for skas
00:40 < Pahan> tt
00:40 < david> ok
00:40 < david> makes sense
00:41 < Pahan> I don't quite want to recompile my kernel and/or reboot.
00:41 < david> nod
00:42 < Pahan> Don't have magic datacenter gnomes in my apartment on that coast :P
00:42 < david> heheh
00:43 < Pahan> Does skas make UML incredibly fast?
00:43 < david> yes
00:43 < david> it's is much better than tt
02:39 < mistik1> oh my!
04:15 < BB> morn
05:50 < zlb> is possible to make the fs encrypted?
05:54 < avoozl> you could use an encrypted loopback
05:55 < avoozl> but it usually is pretty pointless as the data can be recovered when the host server gets hacked and the uml is still running anyhow
05:57 < zlb> uhm..
06:40 -!- herb1 [~herb@2Cust156.tnt5.syd2.da.uu.net] has joined #uml
08:00 -!- herb1 [~herb@2Cust156.tnt5.syd2.da.uu.net] has quit [Quit: Client Exiting]
08:04 -!- tp [~posch@asterix.tkn.TU-Berlin.DE] has joined #uml
08:18 -!- thy_ [koubt9am@u-pl7.ms.mff.cuni.cz] has joined #uml
08:49 -!- war- [war@66.45.37.187] has joined #uml
09:13 -!- jdike [~jdike@jdike.solana.com] has joined #uml
09:13 < jdike> hi guys
09:16 -!- war- [war@66.45.37.187] has quit [Quit: brb]
09:16 -!- _slh__ [~chatzilla@68.154.227.191] has quit [Ping timeout: 488 seconds]
09:20 < green> Hi Jeff!
09:23 -!- war- [war@lucidpixels.com] has joined #uml
10:13 -!- trom [trom@pD9E8299A.dip.t-dialin.net] has joined #uml
10:17 < trom> hallo, I have a problem with the naming scheme of the patches, i do not know what matches to what. can anybody help me?
10:33 -!- smcavoy [~smcavoy@207.61.160.163] has joined #uml
10:41 -!- trom [trom@pD9E8299A.dip.t-dialin.net] has left #uml []
10:52 -!- war- [war@lucidpixels.com] has quit [Read error: Connection reset by peer]
11:03 -!- thy_ [koubt9am@u-pl7.ms.mff.cuni.cz] has quit [Remote host closed the connection]
11:04 -!- thy_ [koubt9am@u-pl7.ms.mff.cuni.cz] has joined #uml
12:13 < Getty> its always funny to hear guys about the "unsecurity of UML"....
12:13 < Getty> 2 sentences and they are quite
12:15 < david> Getty: like what?
12:15 < Getty> "its just a usertask - no special rights" "put it into a chroot, then its unhackable"
12:15 < Getty> i mean...
12:15 < Getty> even if the uml coder would be really really stupid, its out of concept impossible to make a security hole out of this
12:18 -!- thy_ [koubt9am@u-pl7.ms.mff.cuni.cz] has quit [Ping timeout: 492 seconds]
12:20 -!- thy_ [koubt9am@u-pl7.ms.mff.cuni.cz] has joined #uml
12:27 < Getty> david: or do you wanted to hear the sentences of the other side? they have none, its only blabla
12:54  * BB wouldnt put his money on it being unhackable ;)
12:55 < Getty> BB: mmhh....
12:55 < Getty> BB: the only way to get out is hacking the chroot
12:56 < tp> there's the skas patch that might introduce holes, if used
12:56 < Getty> BB: (after hacking the UML and trying to get a shell on the chroot which isn't there)
12:56 < tp> also some of the uml helper apps run as root, like uml_net (i heard)
12:56 < Getty> yeah, i always don't talk about skas patch
12:56 < Getty> of course that would be security hole
12:56 < Getty> tp: you don't need uml_net
12:56 < Getty> tp: you can setup the network before and then forget about uml_net
12:56 < Getty> uml_net is only some kind of helper application to do the dirty work
12:56 < david> the worst skas can do is cause a DOS against the host by eating fds
12:56 < tp> yes, i do it without uml_net too
12:57 < Getty> so at all i think UML is really really secure ;)
12:57 < Getty> hey stupid question beside
12:57 < Getty> what about kernel patches over UML patches?
12:58 < tp> i guess, some will work, and some won't ;-)
12:58 < david> Getty: how do you mean?
12:58 < Getty> i thought about this grsecurity patches or wolk or something combine with UML (not for myself, only for known)
12:58 < david> Getty: I have a pretty customised kernel
12:58 < david> Getty: anything which is arch specific won't work without hacking
12:58 < david> Getty: I use stuff like LVM, ebtables, etc
12:58 < Getty> ah ok...
12:58 < david> uh, LVS
12:58 < david> not LVM
12:59 < Getty> this arch specific is the thing i not thought about...
12:59 < Getty> (that was really bad english ;) )
12:59 < david> Getty: most of the patches I applied are networking
12:59 < david> or non-arch cleanups
13:00 < Getty> btw.. what about this crazy idea with the UML ontop of systems with linux-comp. modes?
13:00 < Getty> solaris, some bsd and so? :) some tests around there? :)
13:00 < david> Getty: no idea
13:01 < david> Getty: probably best to port it to those systems
13:01 < Getty> ooohhh...
13:01 < Getty> you think that is possible?
13:01 < Getty> UserModeLinux for BSD for example as real BSD application?
13:01 < jdike> Getty: they would have to be *extremely* compatible
13:01 < jdike> Getty: and afaik, none is
13:02 < Getty> yeah i only thought about the binary-compatibility modes in this other OS
13:02 < david> Getty: you can, in theory, port UML to non-Linux platforms (BSD, Win32, etc) assuming it's ia32
13:02 < mistik1> hey guys
13:02 < david> mistik1: hello
13:02 < david> jdike: hello
13:02 < mistik1> jdike: thank you
13:03 < jdike> hi david
13:03 < jdike> david: In theory, you can port UML to any platform, IA-32 or not
13:04 < david> jdike: right, but I assume porting UML to another ia32 platform would be easier than porting it to, say, Sparc
13:04 < david> or am I not paying attention
13:04 < jdike> david: Well, porting to foo/x86 and porting to Linux/bar are two different things, and you can't say a priori which is easier
13:04 < Getty> hehe
13:04 < Getty> i go home now
13:04 < Getty> bye guys
13:06 < david> jdike: true
13:09 < caker> I wonder where people are getting the idea the the mailing list is for general linux questions
13:10 < david> caker: I've not seen anything on the uml list for a while
13:11 < caker> david: there's been activity on uml-user for the past few days - not getting them?
13:12 < jdike> caker: there's just ModemMan and this last one
13:12 < jdike> caker: that's hardly a trend
13:12 < caker> jdike: i agree, just wonder how they are making that leap
13:12 < jdike> caker: yeah, me too
13:13 < david> caker: last I got was 6/6
13:13 < jdike> caker: although it may be an infinite monkeys effect
13:13 < david> caker: maybe something thinks they're spam
13:14 < green> hehe
13:14 < jdike> caker: there are enough people out there that any possible conceptual leap will be made by someone
13:14 < caker> jdike: ahh, was just googling for your point :-)
13:27 < mistik1> darm rpm
13:27 < mistik1> so rpm seems to be the most troublesome app inside UML for me
13:28 < mistik1> It does its job but 1 of every 4 jobs is refuses to release the terminal
13:28 < mistik1> so you have to go kill it
13:28 < BB> strange
13:29 < mistik1> have any of you guys come across this behavior?
13:33 < green> mistik1: this is a bug in RH 8 rpm/db/whatever
13:33 < green> it behave s same way on real box too
13:34 < mistik1> ok
13:34 < mistik1> does an update fix this a rebuilding of the db?
13:36 < green> I do not know. I solved it with upgrading to RH 9 ;)
13:36 < mistik1> heh
13:36 < green> rh8 was not very stable distro for me.
13:36 < green> I have much better experiences with rh9
13:36 < mistik1> I have it running a NS in the corner
13:36 < mistik1> that's about all i've done with 8.0
13:37 < caker> What makes linux not use the remaining 4MB of ram I'm giving it? (stupid question, i know)
13:37 < caker> is it initrd support?
13:37 < green> caker: this is a reserve (percntage of your RAM), for urgent kernel needs
13:37 < caker> green: ok .. thought so ..
13:37 < green> It does not use 10M of 2G I gave it ;)
13:38 < caker> green: thanks
13:38 < mistik1> green: I wonder if I could get away with doing what I did from redhat7.2 to 7.3 ;-)
13:39 < mistik1> apt-get dist-upgrade ;)
13:39 < mistik1> It sure didn't work from 7.3 to 8
13:41 < green> mistik1: hm. you mean you have rpm problems in rh 7.3?
13:43  * green actually upgrades between different RH versions with rpm itself, not even rebooting ;)
13:43 < mistik1> green: I mean upgrading as you stated
13:44 < mistik1> I have done so with apt-get for a couple releases
13:44 < mistik1> I dont use redhat kernels anyway so rebooting is a non issue
13:47 < green> well, tihs is not totally seamless and you need to do some hand-dependency-resolving (involves some thinking), but it is certainly possible
13:48 < mistik1> oh, I'm aware of that as i've done it ;)
14:00 < mistik1> hmmm
14:00 < mistik1> time to put the memory on a tmpfs
14:02 < Getty> do you know this: http://freshmeat.net/projects/vserver/ ?
14:03 -!- tp [~posch@asterix.tkn.TU-Berlin.DE] has quit [Quit: Client Exiting]
14:05 < green> yes.
14:05 < Getty> whats the big concept about this?
14:05 < green> they are using this flawed "shared kernel" idea, the same as "freebsd jail", "hspcomplete" and others who hunt for speed
14:05 < Getty> i only see how to use it but i don't see the concept
14:06 < Getty> ah ok
14:06 < green> they patch the kernel so it suddenly have several "security contexsts"
14:06 < Getty> so i can see it as another "UML-patch"?
14:06 < green> each context is isolated from others. Ther is one root context that sees everything
14:06 < Getty> or is it something completly different?
14:06 < green> this is completelly different aproach.
14:06 < green> faster one, but weaker conceptually
14:07 < Getty> so its not a real own system?
14:07 < Getty> canyou explain it with some example?
14:08 < BB> david how did you end up on rockys nntp ?
14:08 < green> no. It is just assume each process have "machine id", so each time you do something, kernel checks your machine id and then determines what you are allowed to do based on this
14:08 < green> there are different implementations of this
14:09 < Getty> so its still using binaries of the host?
14:09 < Getty> did i understand right?
14:09 < Getty> (or other way: the binaries must run on the host kernel)
14:10 < green> Getty: yes, everything lives on the host, only kernel prevents some processes to be visible by others, some resources are invisible. Kind of enhanced chroot
14:10 < green> you break out of this "chroot" (e.g. exploit a kernel bug) and you own entire box immediately
14:12 < BB> muahahaha
14:12 < BB> erm sorry couldnt resist
14:13 < green> no need to say that if you find a way to krash the kernel, everybody dies with kernel
14:13 < green> that's why I disliek this "shaerd kernel" approach ;)
14:13 < green> but it is fast.
14:16 < david> BB: migrated there from the futurenet server
14:17 < BB> david ah, just didnt expect to see you there ;)
14:40 < gump> anyone using grsecurity kernel patches on host/uml kernels?
14:40 < Getty> lol
14:40 < Getty> wasn't that a theme some minutes ago? ;)
14:42 < gump> not as far is i canm be bothered to scroll up
14:42 < gump> i saw vserver mentioned thou
14:43 < gump> and their is some similaritys between the two (from waht i understand of both of them)
14:44 < gump> however im mainly looking to use grsecurity to harden the host kernel (even thou im not ineding to have any ports open externaly etc)
14:44 < Getty> no i only asked about if that works.. nothing really details
14:47 < gump> ok . i scrolled up and indeed you did say you had thought about them but i see no more relavent discusion about grsecuirty aside from david saying he hasnt tried them
14:48 < gump> well once i manage to find some more sca caddys ill give it a shot
15:00 < mistik1> woooo
15:00 < mistik1> http://music.geeksinthehood.net:81/~mistik1/screens/ss_06-10-2003.png
15:17 < gump> windowmaker =]
15:18 -!- Phazeman [~Phazeman@217.132.15.27] has joined #uml
15:18 < Phazeman> mistik1: you here ?
15:21 -!- Phazeman [~Phazeman@217.132.15.27] has left #uml []
15:28 < Getty> mistik1: you called your system "localhost"?
15:34 < david> gump: the grsecurity patches do not work with UML kernels
15:34 < david> gump: no idea if they'll work on a host, but they may block ptrace which UML needs
15:39  * Getty is thinking about.. why someone needs grsecurity in the uml? ;)
15:39 < Getty> anyway i'm out for some minutes
15:53 < david> Getty: to stop someone rooting your UML
16:14 < gump> cheers david
16:15 < gump> as i said once i get some sca caddys and my 1u running uml ill try grsecuirty on the host kernel
16:17 < gump> you dont have to block ptrace (thou it is one of the common uses for grsecuirty) and still make use of acl
16:20 < caker> anyone use screen to provide access to UML's consoles?
16:28 < david> caker: yes
16:28 < caker> david: I'd like to remove the escape and defescape chars to make it secure, but it's not honoring it .. any suggestions
16:30 < david> caker: hrm, I just made a screenrc to take out a load of commands to get them out to a shell
16:31 < caker> david: Did you have to leave the escape char enabled, and then just hope you cought all the sub-commands?
16:32 < david> caker: yep
16:32 < caker> argh
17:14 -!- kara [~kara@64.240.156.226] has joined #uml
17:15 -!- radical [~kernet@202.88.184.39] has joined #uml
17:30 -!- avoozl is now known as avoozzzl
18:14 -!- thy_ [koubt9am@u-pl7.ms.mff.cuni.cz] has quit [Remote host closed the connection]
18:22 -!- smcavoy [~smcavoy@207.61.160.163] has quit [Remote host closed the connection]
18:25 -!- trom [dankbar@217.232.22.242] has joined #uml
18:26 < trom> Hi,
18:26 < trom> after having compiled various kernels with various patches( 2.4.20-1..) I got always, except in one case (patch 2.4.18.1) that Errormessage:
18:26 < trom> (none):/usr/src/uml18.50/kernel-source-2.4.18# ./linux ubd0=/root/Debian-3.0r0.ext2 debug
18:26 < trom> xterm: Can't execvp /usr/lib/uml/port-helper: No such file or directory
18:26 < trom> (none):/usr/src/uml18.50/kernel-source-2.4.18# 
18:26 < trom> Can anybody help me?
18:26 < trom> installing uml-utilities:
18:26 < trom> <trom> I installed it, now the console-window shortly blinks and then vanishes and nothing. ? Maybe compiling it once again? Why did it work this one time with 2.4.18.1?
18:28 < trom> Do anyone know a solution?
18:28 < caker> trom: try appending "con=null con0=fd:0,fd:1"
18:31 < trom> ./linux ubd0=/root/Debian-3.0r0.ext2 debug con=null con0=fd:0,fd:1
18:31 < caker> yup
18:31 < trom> the same
18:32 < caker> did you compile this kernel?
18:32 < trom> now?
18:32 < trom> (:
18:32 < caker> did you compile it yourself, or download it
18:32 < trom> yes 
18:32 < caker> Yes?
18:32 < caker> which
18:33 < trom> 2.4.18 with patch 2.4.18.50
18:33 < trom> 2.4.18-50
18:34 < trom> and others 2.4.20 with to patches each one
18:34 < caker> what's the output of grep CON_ .config
18:35 < caker> err, grep CHAN .config
18:36 < trom> CONFIG_FD_CHAN=y
18:36 < trom> CONFIG_NULL_CHAN=y
18:36 < trom> CONFIG_PORT_CHAN=y
18:36 < trom> CONFIG_PTY_CHAN=y
18:36 < trom> CONFIG_TTY_CHAN=y
18:36 < trom> CONFIG_XTERM_CHAN=y
18:36 < trom> CONFIG_CON_ZERO_CHAN="fd:0,fd:1"
18:36 < trom> CONFIG_CON_CHAN="xterm"
18:36 < trom> CONFIG_SSL_CHAN="pty"
18:36 < caker> ok that looks good
18:37 < caker> When you boot it, do you see anything?
18:37 < trom> as I said in the beginning : shortly blinking of the xterm.
18:38 < caker> not with con=null con0=fd:0,fd:1
18:38 -!- radical [~kernet@202.88.184.39] has quit [Remote host closed the connection]
18:39 -!- radical [~kernet@202.88.184.39] has joined #uml
18:40 < trom> xterm: Can't execvp /usr/lib/uml/port-helper: No such file or directory
18:40 < trom> excuse
18:40 -!- radical [~kernet@202.88.184.39] has quit [Read error: Connection reset by peer]
18:41 < trom> I got this message before i installed uml-utilities, now with or without that line only I get the blinking.
18:41 < caker> http://user-mode-linux.sourceforge.net/switches.html: read about the debug switch
18:43 < trom> Starting periodic command scheduler: cron.
18:43 < trom> Seg fault in signals
18:43 < trom> leaving out debug
18:43 -!- kara [~kara@64.240.156.226] has quit [Read error: Connection reset by peer]
18:43 < caker> might be old bugs, try the newest patches
18:44 < trom> I thought so, and I did it with the same symptoms.
18:46 < trom> 2.4.20 with patch 2.4.20.1 and 2.4.20.6
18:46 < caker> trom: strange
18:47 < trom> only one version functioned 2.4.18 with 2.4.28-1
18:47 < trom> 2.4.18-1
18:47 < caker> do you run skas on the host?"
18:48 < trom> I dont know what it is, I only followed the instructions in the howtos.
18:49 < caker> skas is a patch for the host kernel, it lets UML run better/faster/stronger
18:49 < caker> but it really shouldn't segfault
18:49 < caker> like you're seeing
18:52 < trom> so?
18:52 < caker> so, that's all I can help you with, I'm afraid
18:56 < mistik1> darnit
18:57 < mistik1> I cant find it now and did not bookmark the page, I found some info on the posability of using kernel level autoconfig with DHCP with a uml kernel
18:58 < trom> caker: you helped me, thanx, gonna compile it once again and then post to the list, its late here already, cu
18:59 < caker> trom: no problem -- good luck!
19:23 -!- trom [dankbar@217.232.22.242] has quit [Quit: using sirc version 2.211+KSIRC/1.1]
19:39 -!- Run [carlo@node-d-5886.a2000.nl] has joined #uml
19:40 < Run> hi
19:46 < Run> Linux version 2.5.69-1um (root@ansset) (gcc version 3.2.3 20030422 (Red Hat Linux 3.2.3-4)) #12 Wed Jun 11 01:38:55 CEST 2003
19:46 < Run> On node 0 totalpages: 32768
19:46 < Run> :)
19:47 < Run>  /dev/ubd/disc0:arch/um/drivers/ubd_kern.c:383: spin_lock(arch/um/drivers/ubd_kern.c:a01b6940) already locked by drivers/block/ll_rw_blk.c/1002
19:47 < Run> drivers/block/ll_rw_blk.c:1004: spin_unlock(arch/um/drivers/ubd_kern.c:a01b6940) not locked
19:48  * Run needs the new modules...
19:48 < Run> modprobe: QM_MODULES: Function not implemented
19:48 < BB> i didnt think modules worked in 2.5.x ?
19:48 < jdike> Run: there have been a few complaints about spinlock debugging
19:49 < jdike> Run: haven't got around to looking at them
19:50 < jdike> Run: I've been running with modules disabled in 2.5.{69,70}, so I haven't fixed the module support either
19:51  * BB calls telewest in his pist state
19:54 < Run> jdike: I am *just* about to run it with modules...
19:55 < Run> BB: They definitely work for me on the host.
19:55 < mistik1> jdike: Do you have any idea of somewhere thereis some talk of kernel level autoconfig with UML
19:55 < mistik1> kernel level dhcp
19:59 < jdike> Run: you need to pull in the alternate instruction stuff from i386
19:59 < BB> i would have expected dhcp to be the same, once you have the bridge, or whatever networking the same
19:59 < jdike> Run: shouldn't be a big deal, just haven't done it yet
19:59 < jdike> mistik1: nothing really to do with UML
20:00 < jdike> mistik1: and dhcp is getting pulled out of the kernel anyway
20:00 < BB> jdike is bootp staying?
20:01 < jdike> BB: that stuff is all slated for removal
20:01 < jdike> BB: that's what the early userspace stuff is about
20:01 < BB> jdike hmm so rootnfs isnt likely to work anymore?
20:02 < jdike> BB: it'll work fine
20:02 < jdike> BB: just it's getting pushed into userspace
20:02 < BB> jdike that sounds like a right pain, i liked the everything in kernel solution of booting client machines
20:03 < jdike> BB: everyone else hates it
20:03 < jdike> BB: what's happening is that there will be an initramfs image attached to the kernel binary
20:03 < BB> jdike awe, guess they dont like the ease of bootp to machine installing
20:04 < jdike> BB: that gets mounted and run, and it does whatever setup is needed to make the machine bootable
20:04 < BB> jdike so you would need to do bootp/dhcp then mount the nfs ala initrd i guess
20:04 < jdike> BB: i.e. load the root filesystem/device modules, dhcp, bootp, etc
20:05 < BB> gah, i dont use initrd as its a pain in the ass... i guess i should start to learn :/
20:05 < jdike> BB: initramfs, not initrf
20:05 < Run> Something is wrong with the xterm signals...
20:05 < jdike> initrd
20:06 < jdike> BB: plus, it's dead simple
20:06 < BB> jdike yeah, but its basicaly the same principal i would guess
20:06 < Run> The commandline gets screwed when wrapping around, ... perhaps the resize signal is not propagated correctly?
20:06 < BB> I havnt followed anything > 2.4 so i guess i'm rather behind
20:07 < caker> Run: what distro are you running?
20:07 < Run> Actually, I mean the console... the original term I started uml in.
20:07 < Run> caker: redhat 9
20:08 < Run> <-- Carlo Wood.  I mail -devel a few times.
20:08 < Run> mailed even
20:08 < caker> Run: I had to modify /etc/sysconfig/i18n and change to LANG="en_US"
20:08 < mistik1> this is all very interesting
20:09 < Run> caker: that file doesn't exist on the uml
20:09 < Run> caker: You mean on the host?
20:09 < caker> Run: the uml machine - you might be seeing a different problem
20:10 < Run> Weird...
20:10 < Run> >rpm -qf /etc/sysconfig/i18n
20:10 < Run> file /etc/sysconfig/i18n is not owned by any package
20:11 < Run> redhat 8 idem... more weird.
20:11 < Run> On the host anyway, I already have LANG="en_US"
20:12 < caker> Run: i had issues where my term didn't support UTF-8, that's why I asked :-)
20:27 < Run> Ok, modules work.
20:29 < jdike> Run: 2.5.70?
20:32 < Run> jdike: No, the host is.  But the uml is 2.5.69.
20:33 < Run> caker: It is related to my bash prompt.   When I set a simple prompt, it works.
20:33 < jdike> Run: so, modules now work in UML or on the host?
20:33 < Run> jdike: both - I just made it work in UML.
20:33 < Run> No problems whatsoever there.
20:34 < Run> Not more than what I had getting it to work on the host that is ;)
20:34  * Run tries a prompt without colors.
20:38 < Run> Yeah, its just the colors in the prompt that don't work.
20:38 < Run> That probably means that bash can't figure out the width of the screen...
20:39 < Run> (color codes take no width)
20:40  * Run tries copying 'resize' from X... (I didn't install X).
20:41 < Run> Hmm, that works though...
20:41 < Run> UML:/usr/src>ls /usr/local/bin
20:41 < Run> UML:/usr/src>cp /mnt/ansset/usr/X11R6/bin/resize /usr/local/bin
20:41 < Run> UML:/usr/src>resize
20:41 < Run> COLUMNS=125;
20:41 < Run> LINES=29;
20:45 < Run> Ah, I got it to work correctly.
20:46 < Run> It was my fault - I had added the "UML:" string inside a [], which mark characters that should not be counted for the width.  Therefore bash got confused about the warp point.
20:46 < Run> wrap, even
20:47 < Run> resizing also just works... Ok, what next.
21:07 -!- jdike [~jdike@jdike.solana.com] has quit [Ping timeout: 492 seconds]
21:27 < Run> Is this a normal full backtrace?
21:27 < Run> (gdb) bt
21:27 < Run> #0  0x400c80a5 in nanosleep () from /lib/i686/libc.so.6
21:27 < Run> #1  0xa001ceda in idle_sleep (secs=-4) at arch/um/kernel/time.c:115
21:27 < Run> Thats all I get.  Is there missing something?
21:28  * Run gets the feeling this aint the kernel
--- Log closed Wed Jun 11 00:00:00 2003