Back to Home / #uml / 2007 / 05 / Prev Day | Next Day
#uml IRC Logs for 2007-05-29

---Logopened Tue May 29 00:00:54 2007
00:20|-|shawarma [] has quit [Remote host closed the connection]
00:21|-|shawarma [] has joined #uml
00:21|-|Netsplit <-> quits: peterz, apic
00:22|-|Netsplit over, joins: peterz
00:25|-|apic [] has joined #uml
01:23|-|newb_prog [] has joined #uml
01:53|-|Ancalagon [] has joined #uml
02:24<shawarma>Is the UML in the 2.6.22 kernel tree not supposed to work?
02:28<shawarma>Well... Of course, it's *supposed* to work, but is it also expected to work?
02:36|-|motp [~motp@] has joined #uml
04:00|-|newb_prog [] has quit [Quit: Leaving]
04:28|-|motp [~motp@] has quit [Quit: Leaving]
07:16|-|krau [~cktakahas@] has quit [Ping timeout: 480 seconds]
07:22|-|krau [~cktakahas@] has joined #uml
07:34|-|alb [] has joined #uml
07:37|-|albertito [] has quit [Ping timeout: 480 seconds]
08:26<dgraves>shawarma: i don't see why not.
08:26<dgraves>what's your issue?
08:26<shawarma>I've been debugging it a bit.
08:26<shawarma>I have a backtrace somewhere.. Lemme find it.
08:57|-|linux__alien [~linux__al@] has joined #uml
08:58<linux__alien>has anyone tried installing GNU Zebra in UML?
08:58<linux__alien>i am not able to do it . could someone guide me please?
09:01|-|da-x_work_ [] has joined #uml
09:02<dgraves>linux__alien: what's the error and your procedure? what's different about gnu zebra? it should install like most other distros...
09:04<linux__alien>dgraves, Thanks for replying. I ve installed GNU Zebra in Host kernel and i am able to find the binaries in /usr/loca/sbin but i am not able to access it from UML so how do i install it in UML
09:10<linux__alien>dgraves, any suggestions please?
09:15<peterz>linux__alien: why do you install it on the host if you need it in the guest?
09:16<linux__alien>peterz, i thought i can access that from UML and morever didnt know how to install it in UML . I ve the tar package in my host kernel placed in my desktop folder how do i copy it to UML so that when i boot into UML It becomes visible
09:16<linux__alien>i dont know that
09:17<linux__alien>ve a doubt with GTK+ can i develop applications using C as well as C++ as well as Ruby ?
09:17<linux__alien>is that possible?
09:17<peterz>linux__alien: if you've got networking working you can use any remote cpy/fs otherwise possible
09:18<peterz>linux__alien: that is rather off-topic here, but yes, there are a lot of language bindings for gtk
09:18<linux__alien>peterz, sorry instead of posting it there i posted it here
09:19<linux__alien>peterz, i ve created two interfaces and assigned two addresses and able to ping i ve created two virtual interfaces
09:19<linux__alien>but i am not able to understand how to copy it from my desktop folder in host to UML
09:19<peterz>linux__alien: can you use ssh or nfs from host to guest?
09:20<linux__alien>peterz, let me check one second
09:22<peterz>(/me notes that one second has long since passed)
09:23<linux__alien>:-) kindly just hold on i badly need it for my project . I am working on routing i need to test my code with the routing protocol .
09:23|-|jdike [] has joined #uml
09:23<jdike>hi guys
09:23<linux__alien>hi jdike
09:25<linux__alien>peterz, i am not able to ping my virtuall created interface address from host
09:25<peterz>yo jdike
09:25<peterz>linux__alien: well, get that fixed first
09:27<linux__alien>peterz, ping is working inside UML but not working from the host. Is there anything that i am missing ?
09:27<peterz>linux__alien: probably; I'd recommend you'd get yourself some howto's and start reading them :-)
09:28<linux__alien>peterz, i read through the HowTo especially in Virtual Networking but its kind of tough :-)
09:28<linux__alien>for me atleast
09:36|-|linux__alien [~linux__al@] has quit [Ping timeout: 480 seconds]
09:43|-|linux__alien [~linux__al@] has joined #uml
09:43<linux__alien>UML worked but my internet went down :-)
09:44<linux__alien>peterz, now i get two new interfaces in my host kernel and ping with those 2 interfaces are working . Those interfaces are tap0 and tap1 and ping is working but when i do a telnet to that IP address it says connection refused
09:45<peterz>make sure you've got a telnet daemon running in the guest
09:45<linux__alien>ssh also fails
09:45<linux__alien>i ve ssh running
09:45<linux__alien>but have a doubt here
09:45<peterz>can you ssh to localhost in the guest?
09:47<linux__alien>ssh -t localhost works in UML
09:47<linux__alien>sorry it does not work
09:47<linux__alien>route add -host dev eth0
09:47<linux__alien>oops sorry
09:47<linux__alien>The authenticity of host 'localhost (' can't be established.
09:47<linux__alien>RSA key fingerprint is 6a:cb:a6:b6:85:4e:b0:4f:a8:68:77:65:bc:e6:3f:54.
09:47<linux__alien>Are you sure you want to continue connecting (yes/no)? yes
09:47<linux__alien>Warning: Permanently added 'localhost' (RSA) to the list of known hosts.
09:47<linux__alien>this is what i get
09:48<jdike>that worked
09:48|-|silug_ [] has joined #uml
09:48<peterz>yeah, looks good
09:48|-|silug_ changed nick to silug
09:48<peterz>now try host to guest
09:48<linux__alien>ssh -t
09:48<linux__alien>ssh: connect to host port 22: Connection refused
09:49<linux__alien>/etc/init.d/sshd status
09:49<linux__alien>sshd (pid 1959) is running...
09:49<linux__alien>thats in UML
09:49<jdike>run ifconfig eth0 inside UML
09:50<linux__alien>eth0 Link encap:Ethernet HWaddr 4E:84:C4:A4:7C:17
09:50<linux__alien> inet addr: Bcast: Mask:
09:50<linux__alien> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
09:50<linux__alien>eth1 is this
09:50<linux__alien>eth1 Link encap:Ethernet HWaddr FA:48:81:A8:6C:2F
09:50<linux__alien> inet addr: Bcast: Mask:
09:50<linux__alien> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
09:50<linux__alien>This is in my Host Kernel
09:50<linux__alien>tap0 Link encap:Ethernet HWaddr 6E:24:57:9B:83:96
09:50<linux__alien> inet addr: Bcast: Mask:
09:50<linux__alien> inet6 addr: fe80::6c24:57ff:fe9b:8396/64 Scope:Link
09:50<linux__alien>tap1 Link encap:Ethernet HWaddr 9E:B0:4E:91:B0:70
09:50<linux__alien> inet addr: Bcast: Mask:
09:50<linux__alien> inet6 addr: fe80::9cb0:4eff:fe91:b070/64 Scope:Link
09:51<jdike>I guess you didn't see the big warnings about not assigning the host tap and UML eth interfaces the same IP
09:51<linux__alien>jdike, hmm...actually i had ifcfg-eth0 file which had a different address but then i reconfigured it after booting inside UML
09:54<linux__alien>now ve changed the ifcfg-eth0 also to the ip which i propose to give upon reboot. its rebooting
09:54<jdike>all you needed to do was change the IP
09:56<linux__alien>jdike, yes true so now gave an IP address for eth0 and for eth1 but when booting it said an other host is using the same IP . an interface by name tap0 and tap1 in host uses that and it gets automaticaly created so now it did not configure an ip for eth0 and eth1. What do i do
09:56<linux__alien>i ll have to give the same IP as tap0 and tap 1 right?
09:56<linux__alien>for eth0 and eth1 in UML right?
09:57<jdike>isn't that what I just said you shouldn't do?
09:57<linux__alien>jdike, oh sorry now saw the warning
09:59<linux__alien>jdike, ok now ve configured eth0 and eth1 in UML with and and my host kernel has tap0 and tap1 with and
10:00<linux__alien>now have a doubt how does tap0 have a different IP . Now ping works to tap0 but not to eth0 so i ll have to add a static route in host is it ?
10:00<linux__alien>for ping to succeed ?
10:01<jdike>you're the routing expert
10:01<jdike>you tell me
10:07[~]dgraves wonders if this is an exercise for the reader.
10:09<jdike>someone who is setting up an environment for playing with Zebra and BGP shouldn't be asking questions about setting up static routes
10:13<linux__alien>jdike, the problem is the mask portion. i ve a Class A IP address and its taken 8 as the net mask but while adding a route it takes it as 24
10:13<linux__alien>and hence ping does not work
10:13<linux__alien>infact it takes it as 32
10:13<linux__alien>the net mask portion
10:23|-|Netsplit <-> quits: silug, Urgleflogue, VS_ChanLog, tasaro, Electric1lf, flatronf700B, besonen_mobile, krau, fo0bar, linux__alien, (+16 more, use /NETSPLIT to show all of them)
10:24|-|Netsplit over, joins: VS_ChanLog, besonen_mobile, flatronf700B, caker, SNy, fo0bar, linbot, tasaro, Electric1lf, weasel (+15 more)
10:25<linux__alien>in UML also ve changed the interfaces eth0 and eth1 to 24 Bit Mask
10:25<linux__alien>and changed the tap0 and tap1 also to 24 bit Netmask
10:25<linux__alien>i mean in the host
10:27<shawarma>dgraves: Sorry I ran off like that. I've tried compiling uml from the Ubuntu kernel tree (I'm making an ubuntu package of uml), but it seems no matter which options I pass it, it just gives me this output:
10:28<linux__alien>any ideas jdike ?
10:28<jdike>shawarma, set your coredump ulimit to unlimited and run it again
10:29<jdike>then gdb the core file and get a backtrace
10:31<linux__alien>jdike, even 8 bit netmask does not work
10:35<linux__alien>any help ?
10:38<linux__alien>am i missing something?
10:41|-|kokoko1 [~Slacker@] has joined #uml
10:42<jdike>you've looked at your routes and addresses and made sure they all say what you want?
10:42<shawarma>jdike: I'm fetching the source right now, so I can get a bit of context..
10:43<jdike>panic (fmt=0x81f8348 "Out of low memory")
10:43<shawarma>What does that mean in that context?
10:44<shawarma>uml shouldn't need lowme from the host?
10:44<jdike>it means that UML doesn't have enough memory
10:44<jdike>what was your command line?
10:46<jdike>disable CONFIG_HIGHMEM
10:46|-|the_hydra [~mulyadi@] has joined #uml
10:46<shawarma>In that particular case it was just "linux ubd0=uml.img"
10:46<shawarma>jdike: I'll try that.
10:46<jdike>or give it more memory
10:46<shawarma># CONFIG_HIGHMEM is not set
10:48<jdike>OK, I was misreading the ifdefs there
10:49<shawarma>jdike: Yes?
10:49<jdike># free
10:49<jdike> total used free shared buffers cached
10:49<jdike>Mem: 30140 29040 1100 0 1756 11608
10:49<jdike>that's FC6 booted on 32M
10:51<shawarma>I'm not sure, what you're trying to tell me :)
10:52<jdike>i.e. it WorksForMe (TM)
10:53|-|ram [] has quit [Read error: Operation timed out]
10:53<jdike>what happens if you boot it with 128M or something?
10:53<shawarma>Compiled from a 2.6.22 kernel tree?
10:54<jdike>although this code hasn't changed in ages
10:54<shawarma>jdike: Same thing.
10:55<jdike>what's the host?
10:55<shawarma>Same kernel. :)
10:55<shawarma>Ubuntu gutsy.
10:56<jdike>gdb the core file again and
10:56<jdike>printf "%s", log_buf
11:00<linux__alien>yes i checked it i dont think so i ve a problem with the routes
11:01<the_hydra>care to re-tell what's your problem, people/
11:01<linux__alien>if all the ip addresses are in 24 bit netmask i think it should work i believe but it does not believe
11:01<linux__alien>the_hydra, is that for me?
11:02<linux__alien>the_hydra, i am not able to telnet nor ssh to the interfaces that i created in UML from the host kernel
11:02<the_hydra>linux__alien: yup, also for shawarma if he doesn't mind
11:02<shawarma>Sure. gimme a sec.
11:02<the_hydra>shawarma: pk
11:02<the_hydra>grr, i mean ok
11:03<the_hydra>my brain enters slow path...grrrr
11:03<linux__alien>the_hydra, have added static routes but still ping does not work from host kernel to UML
11:03|-|krau [~cktakahas@] has quit [Quit: Leaving]
11:04<the_hydra>linux__alien: first of all, could you ping that interface (the one inside UML)?
11:04<linux__alien>the_hydra, i am able to ping both interfaces that i created inside UML from UML but not from the host
11:05|-|hfb [] has joined #uml
11:05<the_hydra>could you show us your host routing table entries?
11:05<the_hydra>possibly pasted into pastebin or rafb?
11:06<shawarma>In a nutshell: whenever I try to run uml (compiled from the ubuntu kernel tree (2.6.22-rc1)) I get this out: Here's a backtrace: and finally the log_msg
11:06<jdike>nothing new there
11:07<the_hydra>jdike: known bug?
11:11<jdike>a 4096 byte allocation is failing
11:11<linux__alien>jdike, it worked finally but now ve both in the same network and subnet
11:11<linux__alien>jdike, just changed the host address alone for eth0 and eth1 in UML
11:11<shawarma>Could my problem in any way be related to aio?
11:11<the_hydra>linux__alien: IIRC, tap/tun device and interface inside the UML should be on the same subnet
11:11<jdike>shawarma, don't think so
11:11<linux__alien>oh is it didnt know that
11:11<linux__alien>thanks for that info the_hydra
11:11<jdike>there's no I/O happening
11:11<jdike>the_hydra, it's not
11:11|-|krau [~cktakahas@] has joined #uml
11:11<jdike>it'll work for them to be on separate nets as long as the routing is OK
11:11<linux__alien>when i do a ssh its asking for password . but in UML we dont have any password so what password should i give?
11:11<linux__alien>just pressing enter does not work
11:11<shawarma>jdike: Ok. There was a change in struct iocb in the kernel (removed aio_reserved3), which I had to apply to uml as well, but it also didn't work with an earlier 2.6.22 which did not have this change.
11:12<jdike>shawarma, I might need to get my hands on this
11:12<jdike>is the box accessible to the net?
11:13<shawarma>I'm afraid not.
11:13<shawarma>you mean "from the net", right?
11:13<the_hydra>jdike: this one --> __alloc_bootmem_low (size=4096,... ?
11:14<the_hydra>linux__alien: sure no password at all?
11:14<the_hydra>linux__alien: what do you use to logon then?
11:14<jdike>it's got 32M, which is a bit minimal, but I just booted FC6 on 32M, so it's good enough
11:14<shawarma>Before you start doing a lot of things, it can't be because I call it like "linux ubd0=uml.img mem=192M" ? There's nothing wrong with that command line, right?
11:15<shawarma>jdike: Huh? Where do you see the 32M?
11:15<linux__alien>the_hydra, root but now created a new user in UML and using it now it works but now i want to copy a file from my host kernel to UML how do i do it. ftp does not work in UML i believe so how do i copy it
11:15<jdike>multiple places
11:15<jdike>1 - that's the default when you don't specify it
11:15<shawarma>jdike: The same happens when I pass it mem=256. I just tried that.
11:15|-|rocketbomb [~rocketbom@] has joined #uml
11:15<the_hydra>linux__alien: try to loop mount your filesystem image
11:15|-|rocketbomb [~rocketbom@] has left #uml []
11:15<jdike>2 - <7>On node 0 totalpages: 8192
11:16<the_hydra>linux__alien: that is assuming you use non COW image
11:16<linux__alien>the_hydra, how do i do that ? ve not done that at all
11:16<linux__alien>the_hydra, yes i use a non COW image
11:16<linux__alien>hostfs is not built in my UML i believe
11:16<the_hydra>linux__alien: try to read man mount, especially "loop" option
11:16<the_hydra>linux__alien: or yeah, use hostfs
11:17<linux__alien>hostfs does not work
11:17|-|hfb [] has left #uml [Leaving]
11:18<shawarma>jdike: If you have the exact same source and I give you the .config, I'm using, could you see if you could replicate it?
11:18<jdike>shawarma, are you using defconfig?
11:18<the_hydra>jdike: sorry, I just recall it from the way I set up Qemu's tun/tap based net
11:18<shawarma>jdike: no
11:18<jdike>can you try that?
11:18<shawarma>I could. :)
11:19<shawarma>How long are you around for?
11:20<the_hydra>linux__alien: sorry if I can't recall it perfectly, for example : mount -o loop myimage.img /mnt/myimage
11:20<jdike>all day (US/Eastern)
11:20<the_hydra>shawarma: he's always around if not bugged by Intel bosses
11:20<linux__alien>the_hydra, so i ll have to do that in my host kernel is it
11:20<the_hydra>linux__alien: yes, but don't mount your image while it is accesed by UML
11:20<shawarma>jdike: So that's another 4 hours or so?
11:21<jdike>shawarma, probably somewhat longer, but yes
11:21<the_hydra>linux__alien: in othe word, shut down UML first, then mount loop it
11:21<shawarma>jdike: Ok, good. I'll get back to you in a while. It's time for dinner here in CEST. :)
11:21<jdike>OK, take your time
11:22<linux__alien>the_hydra, i ve a doubt i need to copy files from my home directory in the host kernel to UML so cant i mount my home directory from UML wont that be more simple?
11:22<the_hydra>shawarma: don't you think jeff is nice person? give him applause...
11:22[~]shawarma hugs jdike
11:22<jdike>I think I saw "applause", not "hugs" :-)
11:23<shawarma>Deal with it. :)
11:23<the_hydra>linux__alien: sorry I don't understand what you were saying
11:24<linux__alien>the_hydra, i just want to copy files from my home directory (of my host kernel) to UML so that i can compile that package
11:24<the_hydra>linux__alien: yes, and that's one of the way I know
11:25<linux__alien>the_hydra, can you tell me that . i want to compile GNU zebra in UML but the tar file is in the home directory of my host kernel
11:25<jdike>he has networking going, so figuring out the password situation would let him just scp things in
11:25<linux__alien>jdike, ve created a new user also in UML
11:25<shawarma>What's the guest os?
11:26<jdike>try sshing in as the new user then
11:26<jdike>there may be extra problems for root
11:26<linux__alien>ya ok sure did that it works
11:26<linux__alien>so you want me to use scp
11:41<linux__alien>if i give this command from UML scp localhost:zebra-0.95a.tar.gz ./
11:42<linux__alien>it will search this file in the host kernel right? i mean the home directory in the host kernel right?
11:47<jdike>what's localhost?
11:47<linux__alien>jdike, i tried giving the tap0 address and tried doing scp but it says file not found
11:47<linux__alien>scp .
11:48<shawarma>Well, cover me with eggs and flour and bake me for forty minutes...
11:48<shawarma>It works with defconfig.
11:48<jdike>OK, what did you change?
11:48<shawarma>I ditched my old config and went with defconfig.
11:48<shawarma>Now to work out the diff..
11:49<jdike>your old config wasn't an i386 config?
11:49<shawarma>Sure, just not a default one.
11:49<jdike>it was?
11:49<jdike>or was not?
11:49<shawarma>18:49 < shawarma> Sure, just not a default one.
11:50<jdike>the "Sure" was what I wasn't "sure" about
11:51<shawarma>My old config was a non-default i386 one. There. That should clear out any ambiguity.
11:52<jdike>well, it reverses my understanding
11:52<jdike>you can't use an i386 config on UML
11:52<jdike>any more than you could use it on powerpc or sparc
11:53<linux__alien>jdike, if i ve to compile zebra in UML, i need automake, gcc etc right?
11:53<jdike>I guess
11:53<shawarma>jdike: Oh.. Well, I have different .configs depending on whether I'm compiling on amd64 or i386.
11:53<jdike>I've never done it
11:54<jdike>shawarma, just start with a defconfig and customize that
11:54<jdike>BB fixed the build so it no longer tries to use the host's config
11:54<shawarma>jdike: That's what I did long ago. :) I've been using the same tweaked config for ages, updating it with make oldconfig.
11:55<jdike>but that doesn't protect against someone just dropping in a i386 config
12:01|-|ram [] has joined #uml
12:03|-|ram [] has quit []
12:04|-|ram [] has joined #uml
12:10<the_hydra>gtg all
12:10|-|the_hydra [~mulyadi@] has quit [Quit: using sirc version 2.211+KSIRC/1.2.4]
12:23|-|linux__alien [~linux__al@] has quit [Quit: Leaving]
12:59<kokoko1>ella ella eh eh eh
13:01<caker>feeeeeed me
13:52|-|newb_prog [] has joined #uml
13:55|-|kos_tom [] has joined #uml
16:45|-|krau [~cktakahas@] has quit [Read error: Operation timed out]
17:34|-|kos_tom [] has quit [Remote host closed the connection]
17:44|-|krau [~cktakahas@] has joined #uml
19:14|-|jdike [] has quit [Quit: Leaving]
21:14|-|ram [] has quit [Ping timeout: 480 seconds]
22:59|-|VS_ChanLog [] has left #uml [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #uml
23:08|-|silug [] has quit [Ping timeout: 480 seconds]
23:35|-|alb changed nick to albertito
---Logclosed Wed May 30 00:00:46 2007