Back to Home / #uml / 2007 / 10 / Prev Day | Next Day
#uml IRC Logs for 2007-10-30

---Logopened Tue Oct 30 00:01:24 2007
00:48|-|balbir [~balbir@59.178.49.64] has joined #uml
01:06|-|mgross [~mgross@pool-71-117-236-31.ptldor.fios.verizon.net] has quit [Quit: Leaving]
01:17|-|balbir [~balbir@59.178.49.64] has quit [Read error: Connection reset by peer]
01:34|-|balbir [~balbir@59.178.49.64] has joined #uml
02:53|-|balbir [~balbir@59.178.49.64] has quit [Read error: Operation timed out]
03:08|-|balbir [~balbir@59.178.42.122] has joined #uml
04:11|-|Urgleflogue [~plamen@83.228.65.158] has quit [Ping timeout: 480 seconds]
04:11|-|Urgleflogue [~plamen@83.228.65.158] has joined #uml
04:46|-|ram_ [~ram@pool-71-117-247-93.ptldor.fios.verizon.net] has quit [Ping timeout: 480 seconds]
05:06|-|pieter [~d@dsl-240-149-107.telkomadsl.co.za] has joined #uml
05:06<pieter-#uml->>hi
05:06<pieter-#uml->>could anyone please help me, i'd like to 'export' the host machines' real serial port to be used inside of the uml
05:09[~]pieter #uml pokes Magotari#uml-> pokes Magotari
05:14<Magotari-#uml->>Hey, pieter.
05:15<Magotari-#uml->>I am usually at school at this time, but I stayed at home today.
05:15<Magotari-#uml->>Hmm...
05:15<Magotari-#uml->>Well, I would try to use a redirect of some kind.
05:16<pieter-#uml->>Like how? :)
05:16<pieter-#uml->>hehe
05:16<Magotari-#uml->>You can use ssl=<something> to put UML's ports somewhere. You can then redirect the somewhere to the real serial ports, and from the real ports to somewhere. I think so anyway.
05:16<Magotari-#uml->>Let me check the documentation.
05:17<Magotari-#uml->>http://user-mode-linux.sourceforge.net/old/input.html
05:17<pieter-#uml->>Thanks :]
05:17<Magotari-#uml->>Seems like you can use a file with the ssl=<something>
05:18<Magotari-#uml->>So maybe you can use one of the serial port files?
05:19<pieter-#uml->>so ssl0=tty:/dev/tty0 would make ssl0 on the UML and that would be the host's tty0?
05:19<Magotari-#uml->>I suppose so. Never did it myself, but I think it ought to work.
05:21<pieter-#uml->>I've been trying that, but no device like ssl0 is created inside of the UML
05:21<pieter-#uml->>So thought I might be misunderstanding this
05:23<Magotari-#uml->>Hmm...
05:23<pieter-#uml->>Because, what is ssl0? is that the same as /dev/ttyS0 ?
05:23<pieter-#uml->>*head hurts* :<
05:24<Magotari-#uml->>Hmm... I think it is.
05:24<Magotari-#uml->>An idea here. Try adding the device from mconsole and see what happens then.
05:26<Magotari-#uml->>I am going source code diving to take a look at that code.
05:28<pieter-#uml->>Okay
05:29|-|balbir [~balbir@59.178.42.122] has quit [Ping timeout: 480 seconds]
05:30<Magotari-#uml->>pieter: Anything in your dmesg about serial ports?
05:33<pieter-#uml->>no :<
05:34<Magotari-#uml->>Nothing at all? Very strange, it should say something.
05:36<pieter-#uml->>I'll quickly paste the dmesg
05:36<pieter-#uml->>http://pastie.caboo.se/112280
05:37<Magotari-#uml->>I read a bit, it seems like the device name is ttyS after all, just like on a real machine.
05:37<Magotari-#uml->>So ssl0 would give ttyS0, I guess.
05:38<pieter-#uml->>so doing ssl0=/dev/ttyS0 should have /dev/ttyS0 on the UML be the real ttyS0 port of the system?
05:38<Magotari-#uml->>Initializing software serial port version 1
05:38<Magotari-#uml->>So it does start it up.
05:38<Magotari-#uml->>Yeah, I think so.
05:39<Magotari-#uml->>I am green here, so I cannot promise this is right, but to my eyes this is true.
05:41|-|balbir [~balbir@59.178.52.236] has joined #uml
05:43<pieter-#uml->>Thanks, yeah it also makes abit of sense to me.
05:44<pieter-#uml->>I'll test once I have a device to plug into the serial port.
05:53|-|tyler29 [~tyler@ARennes-257-1-28-109.w81-53.abo.wanadoo.fr] has joined #uml
06:13|-|tyler29 [~tyler@ARennes-257-1-28-109.w81-53.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
06:15|-|balbir [~balbir@59.178.52.236] has quit [Ping timeout: 480 seconds]
06:26|-|tyler29 [~tyler@ARennes-257-1-58-13.w81-53.abo.wanadoo.fr] has joined #uml
06:27|-|balbir [~balbir@59.178.33.21] has joined #uml
06:40|-|ftumch [~James@90.155.42.12] has quit [Quit: Goodbye.]
06:51|-|tyler29 [~tyler@ARennes-257-1-58-13.w81-53.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
07:07|-|tyler29 [~tyler@ARennes-257-1-32-34.w81-53.abo.wanadoo.fr] has joined #uml
07:12|-|balbir [~balbir@59.178.33.21] has quit [Ping timeout: 480 seconds]
07:17|-|tyler29 [~tyler@ARennes-257-1-32-34.w81-53.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
07:19|-|ram [~ram@pool-71-117-247-93.ptldor.fios.verizon.net] has joined #uml
07:32|-|tyler29 [~tyler@ARennes-257-1-1-253.w81-250.abo.wanadoo.fr] has joined #uml
07:44|-|tyler29 [~tyler@ARennes-257-1-1-253.w81-250.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
07:51|-|ftumch [~James@james.1ec.aaisp.net.uk] has joined #uml
08:03|-|ftumch [~James@james.1ec.aaisp.net.uk] has quit [Quit: Goodbye.]
08:04|-|ftumch [~James@james.1ec.aaisp.net.uk] has joined #uml
08:26|-|ram [~ram@pool-71-117-247-93.ptldor.fios.verizon.net] has quit [Ping timeout: 480 seconds]
08:44|-|dang [~dang@nemesis.fprintf.net] has quit [Quit: Leaving.]
09:04|-|dang [~dang@aa-redwall.nexthop.com] has joined #uml
09:38|-|tyler29 [~tyler@ARennes-257-1-149-89.w86-210.abo.wanadoo.fr] has joined #uml
10:22|-|flatronf700B [~flatronf7@ns1.clipsalportal.com] has joined #uml
10:24|-|rjbell4 [~bbell@c-75-67-251-249.hsd1.nh.comcast.net] has joined #uml
10:31<rjbell4-#uml->>I'd like to share a "device" between two UML instances, to simulate a SCSI shared device (shared LUN). I found a post by jdike that would seem to indicate I should just share a file on the host (http://osdir.com/ml/uml.user/2003-11/msg00132.html). However, that doesn't seem to work, as uml is attempting to F_SETLK the file and fails.
10:31<rjbell4-#uml->>Any advice?
10:42|-|Netsplit cation.oftc.net <-> unununium.oftc.net quits: Magotari, tyler29, peterz
10:45|-|Netsplit over, joins: tyler29, Magotari, peterz
10:48|-|arun [~arun@yktgi01e0-s5.watson.ibm.com] has joined #uml
10:49|-|Netsplit cation.oftc.net <-> unununium.oftc.net quits: Magotari, tyler29, peterz
10:50|-|Netsplit over, joins: tyler29
10:50|-|Netsplit over, joins: Magotari, peterz
10:51<caker-#uml->>rjbell4: there's a switch to uml, per ubd device to turn that off ...
10:53<rjbell4-#uml->>caker: That would be great. I couldn't find it. Care to share what that switch is? :)
10:54<pieter-#uml->>Magotari: It works! :]
10:55<pieter-#uml->> ssl0=tty:/dev/ttyS0 allows me to use the HOST's ttyS0 by calling ttyS0 inside the UML
10:55<pieter-#uml->>:]
11:04<caker-#uml->>rjbell4: I think it
11:04<caker-#uml->>rjbell4: I think it's "c", but I'm not seeing any documentation for it
11:04[~]caker #uml checks the code#uml-> checks the code
11:05<caker-#uml->>#define OPEN_FLAGS ((struct openflags) { .r = 1, .w = 1, .s = 0, .c = 0, \
11:05<caker-#uml->> .cl = 1 })
11:06|-|pieter [~d@dsl-240-149-107.telkomadsl.co.za] has quit []
11:07<rjbell4-#uml->>caker: I suppose if I dig, I might be able to figure out how to use that nugget of information, though it would be a big help if you could give an example. For instance, I'm using "ubd1=/alt/scratch.img". How should I modify / augment that?
11:08<rjbell4-#uml->>caker: BTW, I appreciate your assistance.
11:08<caker-#uml->>rjbell4: ./linux help:
11:08<caker-#uml->>ubd<n><flags>=<filename>[(:|,)<filename2>]
11:08<caker-#uml->>firstly, I suggest not using ubd0, ubd1, etc... use ubda, ubdb
11:09<caker-#uml->>so, it should be: ubdac=/blah.img
11:09<caker-#uml->>c for cluster, I believe
11:09<caker-#uml->>rjbell4: er, ./linux --help <--
11:11|-|tuxbubling [~omauras@62.80.120.214] has joined #uml
11:11<tuxbubling-#uml->>hello
11:14<tuxbubling-#uml->>i'm trying to boot the FC5 image given as an example on the website. the problem is that i allways get:
11:14<tuxbubling-#uml->>INIT: Id "c0" respawning too fast: disabled for 5 minutes
11:14<tuxbubling-#uml->>INIT: no more processes left in this runlevel
11:14<tuxbubling-#uml->>at start ... i also tried to change guest inittab to vc/0 and have it under /etc/securetty without luck. Any idea what could be wrong?
11:14<tuxbubling-#uml->>i also noted that udev doesn't start correctly:
11:14<tuxbubling-#uml->>Starting udev: /sbin/start_udev: line 202: 782 Bus error /sbin/udevd -d
11:14<tuxbubling-#uml->>/sbin/start_udev: line 202: 783 Bus error /sbin/udevstart
11:14<tuxbubling-#uml->>[FAILED]
11:14<tuxbubling-#uml->>Could this be linked?
11:16<Magotari-#uml->>Hmm...
11:16<Magotari-#uml->>tuxbubling: What is the host?
11:16<tuxbubling-#uml->>gentoo on 2.6.23 kernel
11:17<Magotari-#uml->>And the guest kernel?
11:18<tuxbubling-#uml->>the one from the website: linux-2.6.22-rc2 and 2.6.23.1-linode35 from the precompiled ones found on wiki
11:18<Magotari-#uml->>Hmm... Your CPU architecture?
11:18<tuxbubling-#uml->>Magotari: should the guest be run as root?
11:18<tuxbubling-#uml->>64b
11:18<Magotari-#uml->>Yeah, I thought so.
11:18<tuxbubling-#uml->>x86_643
11:18<tuxbubling-#uml->>mmmh maybe that's it...
11:19<Magotari-#uml->>Usually running 64bit should not be a problem, but...
11:19<Magotari-#uml->>No, this is not a run as root issue.
11:19<rjbell4-#uml->>caker: Out of curiosity, why do you suggest ubda over ubd0?
11:19<tuxbubling-#uml->>well runing it with sudo doesn't help much :(
11:19<Magotari-#uml->>You don't have to use root unless you want hardware access or something like that.
11:19<tuxbubling-#uml->>well yeah
11:19<tuxbubling-#uml->>ok
11:19<Magotari-#uml->>This seems to be a memory related issue.
11:20<Magotari-#uml->>Hmm... Can you compile your own kernel, for a 64b arch?
11:20<Magotari-#uml->>And use a 64b filesystem with it?
11:20<tuxbubling-#uml->>ouch
11:21<tuxbubling-#uml->>would have been fine if i would have been able to run that out of the box
11:21<tuxbubling-#uml->>at least to check out performances :)
11:21<tuxbubling-#uml->>won't have time to build kernel and FS before a couple of days
11:21<Magotari-#uml->>Yeah, it should work right out of the box.
11:22<Magotari-#uml->>I don't know why it does not in your case.
11:22<tuxbubling-#uml->>:(
11:22<rjbell4-#uml->>caker: With 'ubdbc=/alt/scratch.img' it doesn't lock the file. Thanks!
11:22<Magotari-#uml->>A bus error is bad news, so... Yeah. I don't think you should be getting them.
11:22<tuxbubling-#uml->>could i have an important missing kernel parameter?
11:22<Magotari-#uml->>Now, the inittab thingy...
11:22<Magotari-#uml->>No, that is not the case.
11:23<tuxbubling-#uml->>k
11:23<tuxbubling-#uml->>what about the performance numbers? compared to vservers?
11:24<Magotari-#uml->>Hmm...
11:24<Magotari-#uml->>I don't know what vserver is. Do you mean OpenVZ?
11:24<tuxbubling-#uml->>yes another solution like
11:24<tuxbubling-#uml->>quite similar
11:25<Magotari-#uml->>Hmm, UML is about half to 3/4ths the performance of either Xen or OpenVZ. I did not actually measure it, that is what I eyeballed.
11:25<tuxbubling-#uml->>mmmh
11:25<Magotari-#uml->>If you are doing mostly computations, it won't matter, but things which are heavy on system calls are slower.
11:26<tuxbubling-#uml->>i plan to move my http server on UML
11:26<Magotari-#uml->>I see.
11:26<Magotari-#uml->>Networking will be slower too, but not too much.
11:27<tuxbubling-#uml->>my http server is at this time on a slow hardware XP 1800+ and 512MB
11:27<tuxbubling-#uml->>why slower? couldn't i fix a tun iface on my second real interface?
11:28<tuxbubling-#uml->>btw that's not a heavy traffic http server :)
11:28<Magotari-#uml->>phone
11:28<tuxbubling-#uml->>np
11:28<Magotari-#uml->>brb
11:38|-|hfb [~hfb@pool-71-106-219-180.lsanca.dsl-w.verizon.net] has joined #uml
11:44<Magotari-#uml->>tuxbubling: Back.
11:44<Magotari-#uml->>You see, networking is done via syscalls. And all syscalls are slower in UML.
11:45<Magotari-#uml->>If you don't have high traffic then UML will be great for you, it is fairly easy to set up.
11:45<Magotari-#uml->>Just get lighttpd and you should not have a major problem.
11:46<tuxbubling-#uml->>yep yep
11:46<tuxbubling-#uml->>i'm allready using it :)
11:46<tuxbubling-#uml->>love that http server
11:46<tuxbubling-#uml->>:)
11:46<Magotari-#uml->>I personally like hibachi. Geek factor.
11:47<Magotari-#uml->>You can use tuns, yeah. Or bridging, which is preffered by some people.
11:48<Magotari-#uml->>If this is not a big secret, why do you want to use UML? Security?
11:48<tuxbubling-#uml->>the thing is my desktop is so much powerfull compared to my server, i think that having my server virtualised on my desktop would have it faster than on it's own hadrware
11:49<tuxbubling-#uml->>plus removing the real hardware that costs electric power
11:49<Magotari-#uml->>Hmm...
11:49<Magotari-#uml->>You might want to be careful here.
11:49<tuxbubling-#uml->>to what?
11:50<Magotari-#uml->>The virtual machine has maybe 75% of native performance, if put by itself on a server.
11:50<Magotari-#uml->>But if you load it with desktop stuff too, it will decrease because other things are using its time.
11:50<tuxbubling-#uml->>mmmh
11:50<Magotari-#uml->>If you have a superb desktop computer, then it should be fine.
11:51<tuxbubling-#uml->>got a core 2 duo @ 3.2Ghz
11:51<tuxbubling-#uml->>2Gigs of ram that i plan to get to 4
11:51<Magotari-#uml->>Pretty good, I think the migration might be a good idea.
11:51<tuxbubling-#uml->>and a 10k hard drive
11:51<Magotari-#uml->>Back to 1978 again?
11:52<tuxbubling-#uml->>??
11:52<Magotari-#uml->>10k. Small drive. :) Lame joke. I know.
11:52<tuxbubling-#uml->>lol
11:52<tuxbubling-#uml->>didn't get it ^^
11:52<Magotari-#uml->>Don't worry too much about drive speed here. UML should cache data just fine.
11:53<Magotari-#uml->>Don't give it more than 1/2 of your total ram though, or your desktop won't be usable. Both the real and the virtual systems will be sloooow.
11:53<tuxbubling-#uml->>btw i want to have enough performances on my desktop to be able to play games and such while serving php content and dns request on the other interface
11:54<Magotari-#uml->>If UML won't be fast enough, you can take a look at Xen or KVM. Those are open source and free.
11:54<tuxbubling-#uml->>will save me room and electrical cost :)
11:54<tuxbubling-#uml->>yeah i really like xen
11:54<Magotari-#uml->>I cannot stand it. :)
11:54<tuxbubling-#uml->>i use it at work
11:54<tuxbubling-#uml->>but i'm not sure how the kernel performs on a desktop view
11:54<Magotari-#uml->>?
11:55<Magotari-#uml->>UML has no way to view the guest desktop, unless you want X through the network. Xnest and all.
11:55<tuxbubling-#uml->>i mean the xen kernel as a host for a desktop usage
11:56<tuxbubling-#uml->>with compiz-fusion running games etc
11:56<dang-#uml->>xen's big problem is being stuck on ancient kernels...
11:56<tuxbubling-#uml->>yeah too
11:56<Magotari-#uml->>#define ANCIENT
11:56<dang-#uml->>Makes it hard to get current graphics cards working.
11:56<dang-#uml->>2.6.18 and 2.6.20
11:56<dang-#uml->>er graphics drivers...
11:57<Magotari-#uml->>Geez, that is pretty recent. I expected to hear 2.4.x
11:57<dang-#uml->>X and DRI and DRM are moving so fast, that to keep up for gaming, you need current kernels.
11:57<dang-#uml->>Magotari: Ancient w.r.t. X only. :)
11:57<Magotari-#uml->>Right.
11:57<tuxbubling-#uml->>Magotari: 2.6.20 you miss CFS and SLUB
11:57<Magotari-#uml->>Agreed about X and so on.
11:57<tuxbubling-#uml->>:)
11:57<dang-#uml->>If you don't need gaming or desktop bling, they're fine.
11:57<tuxbubling-#uml->>i need gaming :p
11:57<Magotari-#uml->>Why does everyone want SLUB anyway? I hear it is slower...
11:57<dang-#uml->>And tickless...
11:58<tuxbubling-#uml->>and miss lattest sky2 improvements too...
11:58<tuxbubling-#uml->>bot a big deal that's easily backported
11:58<Magotari-#uml->>A nice solution between Xen and UML would be using KVM.
11:58<dang-#uml->>KVM is not ready for primetime...
11:58<tuxbubling-#uml->>well not sure of that
11:59<dang-#uml->>Not even close.
11:59<Magotari-#uml->>No, but more so than lguest.
11:59<dang-#uml->>I've tried it, and it's generally slower than UML for normal usage.
11:59<tuxbubling-#uml->>if i have to run qemu instance to run my vm that's kind of nasty
11:59<dang-#uml->>Sure, but lguest is mostly a proof of concept thing, I think.
11:59<Magotari-#uml->>dang: Really? I have the opposite conclusion. Do you use skas3?
11:59<dang-#uml->>How simple can VM be made.
11:59<dang-#uml->>Magotari: Nope.
12:00<Magotari-#uml->>:/
12:00<Magotari-#uml->>Gotta try to compare again, one day...
12:00<dang-#uml->>26 is the newest KVM I've tried, tho.
12:00<Magotari-#uml->>Quite a bit back then.
12:01<Magotari-#uml->>I will have to try full skas3 one day.
12:01<tuxbubling-#uml->>so here's the way... i want to have to have no noticeable overhead on the host while running the guest without headaches
12:01<Magotari-#uml->>The patches from uml-user got me nearly there, only some parts do work though.
12:01<tuxbubling-#uml->>any proposition accpeted :D
12:01<Magotari-#uml->>Not enough to get skas3 (noprocmm)
12:02<Magotari-#uml->>Go with UML. I think that is the best right now.
12:02<Magotari-#uml->>Stable as hell, quite fast...
12:02<tuxbubling-#uml->>that's what i tought
12:03<Magotari-#uml->>I would only try something else if UML cannot do it.
12:03<tuxbubling-#uml->>Magotari: now help me getting ride of this "INIT: Id "c0" respawning too fast: disabled for 5 minutes" error :p
12:03<dang-#uml->>And it's not especially taxing on the host, especially if you have multi-core
12:03<Magotari-#uml->>tuxbubling: Ok. Give me a second.
12:04<tuxbubling-#uml->>dang: i'm really waiting for tha Q6600 falling under 150€ :)
12:04<tuxbubling-#uml->>s/tha/the/
12:04<Magotari-#uml->>tuxbubling: try adding con=xterm to the UML command line. (I assume you are running X, and no sudo.)
12:05<tuxbubling-#uml->>i'm connected trough ssh to my host
12:05<tuxbubling-#uml->>no xforwarding
12:05<Magotari-#uml->>Erm, no.
12:05<tuxbubling-#uml->>don't think this will work
12:05<Magotari-#uml->>In guest: init 0
12:05<Magotari-#uml->>In host run UML again, with the bit added to the command line.
12:06<tuxbubling-#uml->>well ok
12:06<tuxbubling-#uml->>well i don't see it running :)
12:07<Magotari-#uml->>And I am also assuming you do have xterm installed...
12:07<Magotari-#uml->>But it should show up after a short while.
12:08<tuxbubling-#uml->>show up what?
12:08|-|ram [~ram@bi01p1.co.us.ibm.com] has joined #uml
12:08<tuxbubling-#uml->>as i said, i'm connected to host trough ssh i won't see xterm popin up
12:09<Magotari-#uml->>Oh. I thought you mean guest. :/
12:09<Magotari-#uml->>Sorry about this.
12:09<Magotari-#uml->>Ok, take two then.
12:09<tuxbubling-#uml->>np ;)
12:10<Magotari-#uml->>con=port:9000, and from the host try telnetting to "localhost 9000"
12:11<tuxbubling-#uml->>telnetd: /usr/lib/uml/port-helper: No such file or directory
12:11|-|jdike [~jdike@pool-71-248-190-161.bstnma.fios.verizon.net] has joined #uml
12:11<jdike-#uml->>Hi guys
12:11<tuxbubling-#uml->>hey
12:13<Magotari-#uml->>tuxbubling: Did you install the extra tools for uml?
12:13<tuxbubling-#uml->>hum actually no
12:13<Magotari-#uml->>Yeah... That explains why that one did not work.
12:14<tuxbubling-#uml->>uml-utils isn't it?
12:15<Magotari-#uml->>I don't remember.
12:15<Magotari-#uml->>I am a Gentoo user too, but I just forgot.
12:16<tuxbubling-#uml->>actually can't find them on gentoo O_o
12:16<Magotari-#uml->> sys-apps/usermode-utilities
12:17<tuxbubling-#uml->>yeah found
12:17<tuxbubling-#uml->>what a name...
12:17<Magotari-#uml->>I know. :/
12:17<Magotari-#uml->>Gentoo just ain't what it used to be... *sigh*
12:18<tuxbubling-#uml->>yeah ...
12:19<tuxbubling-#uml->>ok let me retry then
12:20<tuxbubling-#uml->>ok i see the boot
12:20<tuxbubling-#uml->>i see the udev error
12:20<Magotari-#uml->>Yeah, then you shall not see anything. Do the telnet thing at the host.
12:20<tuxbubling-#uml->>i did
12:20<tuxbubling-#uml->>that's what i see in the telnet :)
12:20<tuxbubling-#uml->>same error
12:20<tuxbubling-#uml->>no login :(
12:20<Magotari-#uml->>:/
12:21<Magotari-#uml->>I'm out of ideas. jdike, can you help a bit?
12:21<Magotari-#uml->>Bus error at udev, and no way to log in.
12:21<Magotari-#uml->>Kernel and fs from the website.
12:21<Magotari-#uml->>64bits.
12:21<Magotari-#uml->>Erm, I mean, the architecture of the host is 64bits.
12:21<tuxbubling-#uml->>yes
12:22<tuxbubling-#uml->>also tried to change tty0 to vc/0
12:22<tuxbubling-#uml->>doesn't help
12:23<tuxbubling-#uml->>also created tty0 before this etc
12:23<Magotari-#uml->>Ever try with init=/bin/bash?
12:24<tuxbubling-#uml->>euh no
12:24<Magotari-#uml->>Yeah, without the con=<whatever> and try to invoke bash as init.
12:24<tuxbubling-#uml->>yeah works
12:24<Magotari-#uml->>One step closer then.
12:25<jdike-#uml->>what's the problem?
12:25<Magotari-#uml->>Bus error at udev, and also cannot log in.
12:25<Magotari-#uml->>Bash as init is fine though.
12:26<jdike-#uml->>x86_64?
12:26<Magotari-#uml->>Aye.
12:26<jdike-#uml->>what versions of host and UML?
12:26<Magotari-#uml->>UML kernel from the website.
12:26<Magotari-#uml->>Guest was 2.6.23-gentoo, I think. Right, tuxbubling?
12:26<tuxbubling-#uml->>host is 2.6.23 from kernel.org
12:26<jdike-#uml->>hm
12:27<Magotari-#uml->>Erm. Yeah, host is 23, guest from the website.
12:27<Magotari-#uml->>Damn, too chaotic. Must slow down. :)
12:27<tuxbubling-#uml->>right
12:27<jdike-#uml->>I might need to try to reproduce this
12:28<tuxbubling-#uml->>you want the exact errors?
12:29<jdike-#uml->>32-bit kernel and filesystem?
12:29<tuxbubling-#uml->>for?
12:29|-|tyler29 [~tyler@ARennes-257-1-149-89.w86-210.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
12:29<jdike-#uml->>UML
12:29<tuxbubling-#uml->>yes the ones from the website
12:29<tuxbubling-#uml->>FC5 image
12:35<jdike-#uml->>Starting udev: /sbin/start_udev: line 202: 671 Bus error /sbin/udevd -d
12:36<jdike-#uml->>something like that?
12:36<tuxbubling-#uml->>yes
12:36<jdike-#uml->>followed by no login prompt?
12:36<tuxbubling-#uml->>yes yes
12:37<jdike-#uml->>for now, 64-bit UML and filesystem should be fine
12:37<tuxbubling-#uml->>which one?
12:37<jdike-#uml->>you'll have to build UML from source, but nagafix has 64-bit filesystems available
12:38<tuxbubling-#uml->>oh ok
12:38[~]jdike #uml adds to the todo listy#uml-> adds to the todo listy
12:38<jdike-#uml->>list
12:39<Magotari-#uml->>jdike: Some things which appear at boot don't get entered into dmesg. Instead of printk() they employ nonfatal(). Is this by design, or shall I ++ my todo list?
12:39<Magotari-#uml->>For example, checking for sysemu.
12:42<jdike-#uml->>yes, it is by design
12:42<jdike-#uml->>printk isn't active until later in boot
12:43<jdike-#uml->>so if something bad happens early, you never see the error because it is stuck in the printk buffer and never dumped to the screen
12:43<jdike-#uml->>so early stuff is done with printf
12:45<tuxbubling-#uml->>jdike: thx pointing me that
12:45<Magotari-#uml->>Right. Gotcha.
12:45|-|tyler29 [~tyler@ARennes-257-1-166-50.w86-214.abo.wanadoo.fr] has joined #uml
12:46<caker-#uml->>jdike: hiya
12:47<caker-#uml->>jdike: I noticed that the "c" (cluster?) and "cl) (not sure what that one is) ubd options aren't documented in --help and elsewhere. Just fyi
12:47<jdike-#uml->>c == cluster, right
12:47<jdike-#uml->>disables file locking
12:47<caker-#uml->>nod
12:47<tuxbubling-#uml->>Magotari: thx to you too to have at least tried :)
12:48<jdike-#uml->>there is no 'cl' - they are all single-letter options
12:51<jdike-#uml->>and no 'l' either
12:52<caker-#uml->>ok -- I nabbed that from the OPEN_FLAGS openflags struct
12:55<jdike-#uml->>yeah, that's O_CLOEXEC
12:56<tuxbubling-#uml->>jdike: working nicely with nagafix stuff
12:56<jdike-#uml->>cool
12:57<tuxbubling-#uml->>yes :)
12:58<tuxbubling-#uml->>now time to migrate the server :)
12:58<tuxbubling-#uml->>just a question
12:58<tuxbubling-#uml->>isn't colinux inspired from uml stuff?
12:58<tuxbubling-#uml->>or the contrary?
12:59<jdike-#uml->>it is
12:59<jdike-#uml->>the guy who wrote it ported UML to windows first
13:05<tuxbubling-#uml->>oh ok
13:06<tuxbubling-#uml->>i'm using colinux at work :)
13:09[~]jdike #uml gets some 2.6.25 patches ready for andrew#uml-> gets some 2.6.25 patches ready for andrew
13:10<jdike-#uml->>tomorrow will be a batch of 2.6.24 fixes, I think
13:10<jdike-#uml->>gotta get them patches out
13:10<tuxbubling-#uml->>thx for all buddies
13:10<tuxbubling-#uml->>cya
13:10|-|tuxbubling [~omauras@62.80.120.214] has left #uml []
13:26<rjbell4-#uml->>I'm seeing an error message from gdb when debugging a programming running in uml.
13:27<rjbell4-#uml->>Error while reading shared library symbols:
13:27<rjbell4-#uml->>find_new_threads_callback: cannot get thread info: generic error
13:27<rjbell4-#uml->>gdb then segfaults
13:27<rjbell4-#uml->>Could this be UML related?
13:28<rjbell4-#uml->>gdb seems to work on the host machine. But it could maybe be a filesystem thing too.
13:29<jdike-#uml->>what version of UML?
13:29<jdike-#uml->>and what architecture
13:31<rjbell4-#uml->>x86_64 arch
13:31<rjbell4-#uml->>UML from a semi-recent kernel.org release
13:31<Magotari-#uml->>jdike: My patches are almost ready too. Should not be more than an hour or two now. Depends on the "Real Life disturbances" factor.
13:32<rjbell4-#uml->>Looks like circa October 2nd
13:33<jdike-#uml->>turn off the phone, close the door, close the curtains, ignore the doorbell
13:33<jdike-#uml->>rjbell4, can you paste a gdb session where this happens?
13:33<jdike-#uml->>on rafb or pastebin
13:35<rjbell4-#uml->>jdike: http://n01se.net/paste/u4N
13:35<jdike-#uml->>can you strace gdb and paste the strace log?
13:35<rjbell4-#uml->>ok
13:36<rjbell4-#uml->>jdike: Any special strace options?
13:36<jdike-#uml->>no
13:36<jdike-#uml->>I generally do -s 1024 if I want to see I/O, but I don't think that's an issue here
13:37<jdike-#uml->>or -ttt if I'm trying to line strace output up with some other log, but that's not a problem here either
13:40<rjbell4-#uml->>http://n01se.net/paste/lqm -- gdb strace
13:43<jdike-#uml->>open("/proc/3017/mem", O_RDONLY) = 9
13:43<jdike-#uml->>pread(9, 0x7f7fdf7370, 1728, 47255550965456) = -1 EIO (Input/output error)
13:43<jdike-#uml->>looks like there's something wrong with the /dev/mem driver
13:43<jdike-#uml->>ptrace(PTRACE_PEEKTEXT, 3017, 0x2afa8a5436d0, [0]) = -1 EIO (Input/output error)
13:43<rjbell4-#uml->>I was just noticing those.
13:43<jdike-#uml->>and PEEKTEXT just looks like it doesn't work
13:44<jdike-#uml->>these look straightforward
13:45<rjbell4-#uml->>jdike: So you think it's my UML kernel config?
13:45<jdike-#uml->>no
13:46<jdike-#uml->>bugs that need fixing
13:46<rjbell4-#uml->>ah!
13:46<jdike-#uml->>you interested in trying out patches?
13:46<rjbell4-#uml->>jdike: Sure
13:47<jdike-#uml->>OK, give me an email address in case you're not hanging out here, pm it if you want
13:49<rjbell4-#uml->>jdike: pm'ed
13:50<jdike-#uml->>OK, today's problems are on the todo list
13:50<jdike-#uml->>32-bit UML on x86_64, --help, and gdb on x86_64
13:52<rjbell4-#uml->>Heh, that second one was triggered by a question of mine, too. Just trying to keep you busy :)
13:54<jdike-#uml->>Bah
13:54<jdike-#uml->>I'll drop that patch on you just to keep you on your toes
14:00[~]jdike #uml disappears for a bit#uml-> disappears for a bit
14:03|-|tyler29 [~tyler@ARennes-257-1-166-50.w86-214.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
14:04<Magotari-#uml->>Hmmm... Crashme is very lucky today.
14:06<Magotari-#uml->>I don't like it. My patch *could* be the cause... Although I cannot imagine why. Probably just paranoia. I will send it in anyway. This is the "huge negative offset" bug being triggered, old news.
14:14|-|tyler29 [~tyler@ARennes-257-1-139-131.w86-210.abo.wanadoo.fr] has joined #uml
14:59<Magotari-#uml->>jdike: First patch sent in. Only to you, without cc to uml-devel. If I wrote something horrendous, I would rather hear that my code is crap from you instead of from the world.
15:00<Magotari-#uml->>I can resend it to uml-devel later.
15:00<Magotari-#uml->>Now, the second one... Gotta work on that a bit.
15:28<Magotari-#uml->>It builds...
15:30<Magotari-#uml->>... boots ...
15:31<Magotari-#uml->>Yeah, seems to work.
15:48<jdike-#uml->>back
15:52<jdike-#uml->>(&new)->sa_handler
15:53<jdike-#uml->>the usual way to spell that is new.sa_handler
15:54<jdike-#uml->>and signal is OK to use here - you don't need the generality of sigaction
15:54<jdike-#uml->>otherwise looks OK
15:56<ds2-#uml->>*(&new).sa_handler ;)
15:57<Magotari-#uml->>Woops. Sorry. Do you want another revision?
15:57<jdike-#uml->>nope, I'll fix it up
15:58<Magotari-#uml->>Thanks a lot. I am about to submit the second one.
15:58<jdike-#uml->>That should be easier
15:58<Magotari-#uml->>Yeah, I am just testing it.
15:59<Magotari-#uml->>I made the first one in such a way that the second one is only removals of a contigous chunk.
16:15<Magotari-#uml->>Ugh. I forgot "uml:" in my patch subject. Sorry about that. After spending two hours on just grinding it into proper shape for submission, I missed the obivous.
16:15<Magotari-#uml->>At least I won't make that mistake with the second one.
16:28<Magotari-#uml->>Sent.
16:29<Magotari-#uml->>I found more freshly unused code to axe, and I did. So the idea of just single chunk of a patch died quietly.
16:43<Magotari-#uml->>Yawn. No patches to work on. This is boring. Time to find some bugs to patch.
16:46|-|dgraves [~agraves@inet-netcache2-o.oracle.com] has quit [Server closed connection]
16:46|-|tyler29 [~tyler@ARennes-257-1-139-131.w86-210.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
16:46|-|dgraves [~agraves@inet-netcache2-o.oracle.com] has joined #uml
16:58|-|dang [~dang@aa-redwall.nexthop.com] has quit [Quit: Leaving.]
17:04|-|tyler29 [~tyler@ARennes-257-1-133-13.w86-210.abo.wanadoo.fr] has joined #uml
17:23|-|tyler29 [~tyler@ARennes-257-1-133-13.w86-210.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
17:23|-|kos_tom [~thomas@humanoidz.org] has joined #uml
17:35|-|tyler29 [~tyler@ARennes-257-1-122-178.w86-210.abo.wanadoo.fr] has joined #uml
17:52|-|arun [~arun@ool-44c6b6b1.dyn.optonline.net] has joined #uml
18:52|-|tyler29 [~tyler@ARennes-257-1-122-178.w86-210.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
18:55|-|jdike [~jdike@pool-71-248-190-161.bstnma.fios.verizon.net] has quit [Quit: Leaving]
18:59|-|kos_tom [~thomas@humanoidz.org] has quit [Remote host closed the connection]
19:10|-|tyler29 [~tyler@ARennes-257-1-92-118.w81-48.abo.wanadoo.fr] has joined #uml
19:19|-|tyler29 [~tyler@ARennes-257-1-92-118.w81-48.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
19:47|-|weasel [~weasel@asteria.debian.or.at] has joined #uml
19:58|-|hfb [~hfb@pool-71-106-219-180.lsanca.dsl-w.verizon.net] has quit [Quit: Leaving]
20:51|-|ram [~ram@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds]
23:08|-|Nem^ [~Nem@dslb-084-056-234-146.pools.arcor-ip.net] has quit [Ping timeout: 480 seconds]
23:17|-|Nem^ [~Nem@dslb-084-056-250-084.pools.arcor-ip.net] has joined #uml
23:59|-|VS_ChanLog [~stats@ns.theshore.net] has left #uml [Rotating Logs]
23:59|-|VS_ChanLog [~stats@ns.theshore.net] has joined #uml
---Logclosed Wed Oct 31 00:00:23 2007