Back to Home / #uml / 2007 / 03 / Prev Day | Next Day
#uml IRC Logs for 2007-03-29

---Logopened Thu Mar 29 00:00:39 2007
00:20|-|cmantito [] has joined #uml
01:33|-|ram [] has quit [Ping timeout: 480 seconds]
04:22|-|tyler [] has joined #uml
05:18|-|tyler [] has quit [Ping timeout: 480 seconds]
05:32|-|tyler [] has joined #uml
05:36|-|jeroenh [] has joined #uml
05:36<jeroenh>hi all
05:36<jeroenh>Where can I get the latest version of the uml-utilities ?
07:13|-|krau [~cktakahas@] has quit [Quit: Leaving]
07:19|-|krau [~cktakahas@] has joined #uml
08:08|-|weasel_ [] has joined #uml
08:08|-|weasel_ changed nick to Guest153
08:09|-|weasel [] has quit [Killed (Guest153 (moo!))]
08:09|-|Guest153 changed nick to weasel
09:18|-|hfb [] has joined #uml
09:18|-|hfb [] has left #uml []
09:22|-|krau [~cktakahas@] has quit [Quit: Leaving]
09:35|-|TBBle [~tbble@] has joined #uml
09:36<TBBle>Is there a way to start uml with eth0 as daemon, such that eth0 gets a randomised MAC address? It currently gets fe:fd:00:00:00:00 which means multiple instances all get the same IP from the dhcp server.
10:00|-|krau [~cktakahas@] has joined #uml
10:03<jeroenh>TBBle: if you use the latest kernel code it should get a random mac address
10:04<jeroenh>or you can assign a mac address yourself at startup
10:04<jeroenh>(on the command line)
10:04|-|jeroenh [] has quit [Quit: jeroenh]
10:25|-|jdike [] has joined #uml
10:25<jdike>Hi guys
10:34<kokoko1>bonjour jdike
10:36|-|krau [~cktakahas@] has quit [Quit: Leaving]
12:04|-|kokoko1 [~Slacker@] has quit [Quit: Leaving]
12:05|-|ram [] has joined #uml
12:11|-|kokoko1 [~Slacker@] has joined #uml
12:12<apic> /away psychotherapie
12:12<apic>oops ;)
12:27|-|tyler [] has quit [Ping timeout: 480 seconds]
12:30<kokoko1>News says "Linux Preinstalled Dell Available Soon"
12:31<kokoko1>here is a joke from /.
12:31<kokoko1>Dell: How can I help you today?
12:31<kokoko1>Me: I bought this computer with Linux on it from you guys, and now I am having problems with X.
12:31<kokoko1>Dell: RTFM, n00b!
12:32<jdike>I bought a Dell with no OS
12:32<jdike>and they tried to sell me extended support on it
12:32<jdike>and I told them there was no way they were going to support this after I put Linux on it
12:33<kokoko1>true, coz they don't know who to spell linux...
12:33<kokoko1>Dell: Hello, can we help you?
12:33<kokoko1>Me: Hi, I am having some problems installing Linux on my new Dell laptop. I need some information about the video card so I can set up X.
12:33<kokoko1>Dell: Ok, umm ... sure ... you said Lyn-ucks?
12:33<kokoko1>Me: Yes?
12:33<kokoko1>Dell: Ok, umm ... can you spell that please?
12:36[~]kokoko1 back to installing updates on FC6 VMs :(
12:41|-|tyler [] has joined #uml
12:50|-|krau [~cktakahas@] has joined #uml
12:57<kokoko1>jdike, I am wondering henrik still in touch with you ?
12:58<kokoko1>He is very active on squid dev. ,userlist and also in #squid on freenod
13:09|-|tyler [] has quit [Ping timeout: 480 seconds]
13:17<jdike>no, I haven't heard from Henrik in quite a while
13:22<baroni>Hi all
13:22<baroni>jdike: I've finished some changes to umlgdb
13:22<baroni>it's working with the 2.6.21-rc1
13:22<baroni>i'm documenting it now
13:22|-|tyler [] has joined #uml
13:23<baroni>but i've changed many things in it
13:23<baroni>i've created an 'umlgdb_dispatcher' and an 'umlgdb_wrapper'
13:24<baroni>maybe it can be useful to others
13:26<baroni>when 'insmod' at the virtual machine, the gdb stops right before the sys_init_module calls module->init , leaving the prompt with a 's[tep]' to the first line of the module being executed, and with its symbols already loaded in the modules list global pointer at the module->module_core location, ready to be debugged
13:33|-|tyler [] has quit [Ping timeout: 480 seconds]
13:33<jdike>did you ever figure out what that 'not a valid object file' error means?
13:36<baroni>i got a lot of gdb errors with libraries being loaded and leaved this behind
13:36<baroni>what it means?
13:37<jdike>no idea
13:37<baroni>the linux kernel file is not a valid object file looked as gdb mistake to me
13:37<jdike>it's a normal ELF binary
13:37<jdike>and you can debug it just fine
13:46|-|waldner [] has joined #uml
13:47|-|baroni [] has quit [Remote host closed the connection]
13:49<waldner>jdike: could you please look at this
13:51<waldner>it puts (<umid>) in xterm title
13:52|-|tyler [] has joined #uml
13:57|-|baroni [] has joined #uml
13:59<jdike>want to send me a signed-off version of that?
13:59<waldner>Ok, how do I sign it?
13:59<jdike>At the bottom put
14:00<jdike>Signed-off-by: Your Name <your email address>
14:00<waldner>ah ok
14:00<jdike>and I'll send it to mainline
14:01<waldner>ok, I'll send it to you ASAP
14:04<kokoko1>wow cool
14:05<jdike>waldner, also a little changlog describing what the patch fixes
14:05<jdike>doesn't need to be a lot in this case coz the patch is almost self-documenting
14:06<waldner>jdike: is it OK to put 1-2 lines describing the fix at the top?
14:07<waldner>sorry but I never did a kernel patch before
14:08<kokoko1>waldner, dude you did surprise boss :P
14:08<jdike>it's almost fine as is
14:08<jdike>just add the changelog and signed-off-by and it'll be perfect
14:11<waldner>does the email need to have a particular subject?
14:11<kokoko1>Signed-off-by: David S. Miller <>
14:11<kokoko1>from kernel changlog :D
14:12<apic>is there any for-dummies-howto to set up a tun/tap tunnel for uml?
14:12<jdike>if it sent that in, I'd get murdered :-)
14:12<apic>the host has multiple ips on eth0.. with ugly aliases like eth0:1
14:12<kokoko1>its just an example for waldner dude
14:12<apic>i want to have 1 ip as a tap dev
14:12<apic>so i can send raw ethernet
14:12<apic>do i need to waste an ip?
14:13<jdike>apic, have you seen
14:13|-|tyler [] has quit [Ping timeout: 480 seconds]
14:13<apic>or does tun/tap work with just 1 ip too?
14:13<apic>yes, i have
14:13<jdike>you can reuse an eth0 IP on the tap device
14:14<waldner>jdike: ok, sent
14:14<waldner>kokoko1: thanks, I also looked on to see how they look like
14:15<waldner>jdike: I used another email address because the one I used last time is going to be closed soon
14:15<kokoko1>waldner, your welcome :)
14:15<waldner>kokoko1: thanks
14:15<kokoko1>waldner, wish you new email id from ibm, intel or oracle for kernel development :)
14:16<waldner>kokoko1: well, I'm not exactly into it...but thanks anyway! :)
14:16<kokoko1>hmm then how this patch come? :)
14:17<kokoko1>mean would you please shed some light on the background for this patch?
14:17<waldner>kokoko1: well, I know a little C
14:18<waldner>kokoko1: "I'm not into it" meant: I do not program for a living
14:18<apic>exactly what i needed :-D
14:18<kokoko1>waldner, right
14:19<waldner>kokoko1: but, for simple things like these, a little tour of the sources can be enough
14:20<kokoko1>yep, so where in .uk ?
14:20<waldner>kokoko1: but I would not probably be able to deal with tougher issues
14:20<waldner>no, I'm in italy
14:20<kokoko1>oh right, stupid whois :-S
14:21<kokoko1>waldner, dude contribute you sound good in program
14:22<waldner>kokoko1: I do not have much time, and I prefer system and network administration
14:22<kokoko1>heh, waldner this is what i do linux sysadmin
14:23<kokoko1>and i am proud administrating 3 host running 13 UML :)
14:24<waldner>kokoko1: basically, I use UML to test things
14:24<kokoko1>waldner, we are using it in production :)
14:24<kokoko1>mail server, jabber server, RT servers, ldap etc
14:24<waldner>kokoko1: yes, I'd love to, if only I was the only one to take decisions...
14:25<kokoko1>This save our boss lot of extra bulks
14:25<waldner>yes, I'm waiting for him to understand that....
14:25<waldner>well, gtg now
14:26<kokoko1>see ya
14:26|-|waldner [] has left #uml []
14:29|-|tyler [] has joined #uml
14:33|-|VS_ChanLog [] has joined #uml
15:15|-|tyler [] has quit [Ping timeout: 480 seconds]
15:15|-|fghj [~tkxue@agp.Stanford.EDU] has joined #uml
15:29|-|tyler [] has joined #uml
15:43|-|tommie [~tommie@] has joined #uml
15:56|-|the_hydra [~mulyadi@] has joined #uml
15:57<the_hydra>hello everybody
16:03<the_hydra>jdike: may I ask something about gdb and printing stack trace?
16:03<jdike>you may ask
16:03<jdike>it may be that no one will answer :-)
16:04<the_hydra>I try to print the stack trace of a binary compiled with -fomit-frame-pointer
16:04<the_hydra>and bt fails when it is more than 1 call depth
16:05<jdike>because the stack trace is non-deterministic without frame pointers
16:05<the_hydra>what I'd like to know is, does gdb employ certain heuristic to guess these frame address?
16:05<jdike>define "fails"
16:06<the_hydra>fails here mean showing wrong return addresses
16:07<jdike>don't know about that
16:07<the_hydra>such as EIP 0x7
16:07<the_hydra>saved EIP, i mean
16:07<jdike>I was just about to propound a theory, but EIP == 0x7 destroys it
16:08<the_hydra>hm, what theory?
16:08<jdike>the theory would have been, it climbs the stack looking for things that look like code pointers
16:08<jdike>and assumes each one is the boundary of a stack frame
16:08<jdike>the kernel stack dumper does something like that
16:09<the_hydra>hm, that could be the way gdb do the heuristic
16:12<jdike>except that 0x7 is not a possible code pointer
16:13<the_hydra>ok thanks for the ideas
16:13<the_hydra>so I guess it's not really easy to print correct trace without frame you said
16:14<jdike>you have to know how big each frame is
16:14<jdike>the code knows, because it built the frame
16:14<jdike>but nothing else does, unless it disassembles the code
16:15<jdike>another heuristic would be to look for a sub $n, %esp at the end of a function
16:15<jdike>and assume that's popping the frame
16:15|-|tommie [~tommie@] has quit [Quit: Leaving]
16:16<the_hydra>bingo, I didn't think about your last idea!
16:19<jdike>but it might turn out that a pop is quicker than a sub for a one word stack, so you'd miss those
16:21<jdike>or maybe add is faster than sub on some CPU
16:21<baroni>jdike, i've updated
16:21<baroni>now a section about 'debugging modules'
16:21<jdike>or maybe gcc knows something about the frame alignment and subtracts 16 with an and
16:22<baroni>and there is the files umlgdb_dispatcher, umlgdb_wrapper and the .gdbinit used
16:22<baroni>now i think it is enoght documented to be used
16:22<jdike>lots of portuguese that I can't read :-)
16:22<baroni>needed to translate to english
16:22<baroni>i'll translate it this weekend
16:23<baroni>i was studing the documentation on uml site, wiki and google, to write this one
16:23<baroni>there is a lot of docs a little bit out of date
16:23<baroni>i'll try to keep this refreshed
16:24<baroni>started as a doc to O.S. students to use uml to study and devel exercices
16:27<caker>!babelize portuguese english Programas no espaço do usuário - disponíveis sob um sistema de arquivos
16:27<linbot>caker: Programs in the space of the user - available under a system of the archives
16:27<baroni>i don't know if the changes made on umlgdb can be ported, since iirc 'gdb' parameter is not anymore used with 2.6.21 kernels (right?), and so i changed the scheme
16:28<baroni>smart bot!
16:30<the_hydra>jdike: sorry was away... thanks again..
16:34<the_hydra>baroni: congratz for your docs
16:36<baroni>hi Mylyadi!
16:36<baroni>thank you :)
16:37<the_hydra>but pls, English version :)
16:38<the_hydra>or, I am forced to drag some Timorese here and interpret them for me :)
16:43<baroni>i'll, this weekend ;)
16:43<the_hydra>looking forward to read that! :)
16:44<baroni>hehe, thanks
16:49|-|krau [~cktakahas@] has quit [Quit: Leaving]
17:03|-|baroni [] has quit [Remote host closed the connection]
17:04|-|fghj [~tkxue@agp.Stanford.EDU] has quit [Ping timeout: 480 seconds]
17:36|-|baroni [] has joined #uml
18:05|-|tyler [] has quit [Remote host closed the connection]
18:16|-|fghj [~tkxue@agp.Stanford.EDU] has joined #uml
18:32|-|fghj [~tkxue@agp.Stanford.EDU] has quit [Quit: Leaving]
19:13|-|the_hydra [~mulyadi@] has quit [Quit: using sirc version 2.211+KSIRC/1.3.12]
19:42|-|jdike [] has quit [Quit: Leaving]
21:28|-|ram [] has quit [Ping timeout: 480 seconds]
21:33|-|fghj [~tkxue@agp.Stanford.EDU] has joined #uml
21:39|-|fghj [~tkxue@agp.Stanford.EDU] has quit [Quit: Leaving]
22:30|-|fghj [~tkxue@agp.Stanford.EDU] has joined #uml
22:45|-|fghj [~tkxue@agp.Stanford.EDU] has quit [Remote host closed the connection]
22:59|-|VS_ChanLog [] has left #uml [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #uml
---Logclosed Fri Mar 30 00:00:21 2007