Back to Home / #uml / 2007 / 01 / Prev Day | Next Day
#uml IRC Logs for 2007-01-03

---Logopened Wed Jan 03 00:00:20 2007
03:38|-|tyler [] has joined #uml
05:37|-|the_hydra [~mulyadi@] has joined #uml
05:55|-|tyler [] has quit [Read error: Connection reset by peer]
05:59|-|tyler [] has joined #uml
06:05|-|tyler [] has quit [Remote host closed the connection]
06:06|-|richardw [] has joined #uml
06:06|-|the_hydra [~mulyadi@] has quit [Quit: using sirc version 2.211+KSIRC/1.3.12]
06:18|-|albertito [] has quit [Ping timeout: 480 seconds]
06:20|-|tyler [] has joined #uml
06:35|-|tyler [] has quit [Ping timeout: 480 seconds]
06:51|-|tyler [] has joined #uml
06:59|-|mkhl [] has joined #uml
07:28|-|richardw [] has quit [Remote host closed the connection]
07:31|-|richardw [] has joined #uml
07:32|-|tyler [] has quit [Ping timeout: 480 seconds]
07:45|-|tyler [] has joined #uml
08:33|-|albertito [] has joined #uml
08:38|-|Newsome [] has joined #uml
08:39|-|richardw_ [] has joined #uml
08:46|-|richardw [] has quit [Ping timeout: 480 seconds]
09:01|-|SiLLeNt [~FFdVv@] has joined #uml
09:43|-|tyler [] has quit [Ping timeout: 480 seconds]
09:50|-|albertito [] has quit [Read error: Connection reset by peer]
09:51|-|albertito [] has joined #uml
09:51|-|da-x [] has quit [Ping timeout: 480 seconds]
09:56|-|tyler [~tyler@] has joined #uml
10:00|-|da-x [] has joined #uml
10:04|-|jdike [~jdike@] has joined #uml
10:04<jdike>Hi guys
10:04|-|kokoko1 [~Slacker@] has joined #uml
10:22|-|tyler [~tyler@] has quit [Remote host closed the connection]
10:30|-|hfb [] has joined #uml
10:36|-|rasix [~jeruk@] has joined #uml
10:49|-|SiLLeNt changed nick to SiLLeNt-00ffzz
10:51|-|hfb [] has left #uml [Leaving]
11:16|-|kos_tom [] has joined #uml
11:41|-|the_hydra [~a_mulyadi@] has joined #uml
11:56|-|the_hydra [~a_mulyadi@] has quit [Ping timeout: 480 seconds]
12:05|-|SiLLeNt-00ffzz [~FFdVv@] has quit [Quit: /raw kill -9 *]
12:13|-|the_hydra [~a_mulyadi@] has joined #uml
13:06|-|richardw_ [] has quit [Remote host closed the connection]
13:36|-|ram [] has joined #uml
13:36<ram>oops. never mind. sorry.
13:41<the_hydra>rasix: hi
13:41|-|richardw [] has joined #uml
13:41<rasix>i'm trying to compile uml kernel for first time
13:41<rasix>i hope it goes well
13:44<the_hydra>good luck
13:44<the_hydra>previously, you just used precompiled binaries?
13:46<rasix>hu uh
13:46<the_hydra>2.4.x one? or 2.6.x?
13:46<rasix>i need some kernel support to get some features running
13:47<the_hydra>like what?
13:49<rasix>sock paket
13:50<rasix>yes .. better try to compile for my own .. and get dirty on it
13:54<the_hydra>hm, I surely miss it, could you explain what sock packet is?
13:55<rasix>when i try to nmap localhost
13:55<rasix>thereis an error ... and the message says that it could be mod for sock packet can not be found
13:56<rasix>dont ask me too much the_hydra.. hehe saya malu disini ada authornya
13:58<the_hydra>on the contrary, it is the right place to ask
13:58<the_hydra>so IMHO no need to be shy or afraid to ask
13:58<rasix>:) .. thanks
13:58<the_hydra>maybe you forgot to compile socket support?
13:59<the_hydra>AFAIK, socket should always there...
13:59<rasix>i thought i grab kernel binary that missing socket support
14:00<the_hydra>or, maybe you put the kernel modules inside the UML root fs
14:00<the_hydra>hence the failure to probe and load certain kernel modules inside UML
14:02<rasix>i mount root_fs under mnt point
14:02<the_hydra>linux --showconfig
14:02<the_hydra>should be helpful
14:02<rasix>and copy modules to mnt/lib/modules
14:03<rasix>ok .. i'll try
14:04<the_hydra>good luck
14:05<the_hydra>observing dmesg message inside UML could be handy too
14:09|-|mkhl [] has quit []
14:11|-|Newsome [] has quit [Quit: Linux: Now with employee pricing!]
14:18|-|richardw [] has quit [Quit: Leaving]
14:20|-|richardw [] has joined #uml
14:38|-|pika [] has joined #uml
14:38<pika>Someone running a 2.6.19 Kernel with skas3?
14:40<rasix>it seems broke when u compile 2.6.19
14:40|-|richardw_ [] has joined #uml
14:41<rasix>i did :D
14:41<pika>Did you apply the 2.6.18 patch on the .19 sources?
14:43<rasix>i didn't :(
14:44<caker>jdike: any more thoughts on ?
14:45<the_hydra>pika: any problem with .19?
14:45<rasix>:) /me noob .. still using old kernel
14:45<caker>jdike: also, curious where that bug is being discussed; or what BB's referring to -- any clues?
14:46<jdike>I don't know what bug he's referring to
14:47<jdike>I haven't really looked at that problem yet
14:47|-|richardw [] has quit [Ping timeout: 480 seconds]
14:49<pika>the_hydra: Nope, some problems with the .18 version. I'ld like to give a self compiled 2.6.19 a try because the linode 2.6.19 works flawlessly. Any idea how to build a 2.6.19 with skas3?
14:51<the_hydra>pika: may I know exactly what kind of error message you saw during compilation?
14:56<pika>the_hydra: I haven't tried yet. See, I've found the skas guest patch for 2.6.18, but not for .19 - is it possible to apply older kernel patches to a 2.6.19 source? If so, how? As usual?
14:57[~]dgraves just waits for skas4
14:57<dgraves>or maybe skas52.
14:58<the_hydra>pika: just try it, with patch --dry-run if you want to play safe
14:58<the_hydra>hi allan
14:58<pika>I'll give it a try and come back later, thanks for now!
14:58|-|pika [] has quit [Quit: [BX] mIRCrap: JUST DON'T DO IT!]
14:58<the_hydra>happy new year 2007
14:59<dgraves>happy ny to you as well. :)
15:04<rasix>i just get error compile the_hydra
15:05<rasix>undeclared "something"
15:05<rasix>it seems use sys-x86_64 instead of i386
15:05<rasix>do i need to try i386? how can i do that?
15:07<the_hydra>are you on x86 32 bit right now?
15:07<the_hydra>how did you do the compilation anyway?
15:08<rasix>i use x86_64
15:08|-|albertito [] has quit [Read error: Connection reset by peer]
15:08<rasix>do i need copy linux-2.6.x to /usr/local/src
15:08<rasix>make menuconfig ARM=um
15:09<rasix>and make ARCH=um
15:09<jdike>don't copy it anywhere
15:09<jdike>and a defconfig before the menuconfig would be a good idea
15:10<rasix>i see
15:13|-|albertito [~net@] has joined #uml
15:14[~]dgraves stops confusing the issue.
15:14<dgraves>happy new year, jdike.
15:14<dgraves>man, i miss having time to hang out with all you guys.
15:15<jdike>dgraves, you too
15:15|-|kos_tom [] has quit [Ping timeout: 480 seconds]
15:15<the_hydra>dgraves: what's the news, allan?
15:15<the_hydra>dgraves: you finally become the new Oracle CTO? ;)
15:17[~]kokoko1 have a headache :(
15:17<the_hydra>kokoko1: please get a rest
15:17<kokoko1>the_hydra, yep leaving ... thanks
15:18<dgraves>the_hydra: please. I'd have hired you already. :)
15:18<dgraves>nah, just running around doing some minor project management and infrastructure.
15:20<the_hydra>dgraves: be patient allant, time will come
15:21<the_hydra>dgraves: anyway, you're not one of the Saddam's executioner, right?
15:22<dgraves>no, i'm not.
15:22<the_hydra>oh i see
15:22<dgraves>i'd have covered him with alfalfa and put him in with rabid bunnies. :)
15:24[~]jdike performs some minor surgery on the ubd driver and it just worked
15:24[~]jdike must have missed something
15:25<the_hydra>another UML tweak, jeff?
15:25<jdike>I'm in the middle of a SMP locking pass over UML
15:26<jdike>I had some locking changes in the ubd driver sitting around, so I decided to split them out
15:26[~]dgraves wonders if we're getting smp? or just smplocks. :)
15:26<jdike>per-device spinlocks and request queues
15:26<jdike>you're getting SMP
15:26<jdike>maybe YOU won't get it, but everyone else will
15:27<dgraves>yay me!
15:27[~]dgraves wishes he could convince his org to upgrade uml versions now.
15:27<dgraves>but 2.6.14 is "working". :(
15:27<dgraves>and they don't want to get ahead of the enterprise linuxes.
15:28<jdike>if(!strcmp(getpwuid(getuid()), "dgraves") ncpus=1
15:29[~]dgraves nukes jdike with leftover eggnog.
15:29<jdike>if(!strcmp(getpwuid(getuid())->pw_name, "dgraves") ncpus=1
15:34<dgraves>good thing i don't have to clean it up, eh?
15:35<the_hydra>cooking an eggnog today, allan?
15:36<dgraves>going out to margarita's though.
15:36<dgraves>food and fun.
15:36<dgraves>i'm on a diet, so no eggnog this christmas.
15:38<the_hydra>you forgot the other "f"
15:38<dgraves>none of them.
15:38<dgraves>jdike has the only hot engineer around here.
15:38<jdike>and I'm keeping her
15:38|-|pika [] has joined #uml
15:39<pika>repetitive 'hi' = rehi :-)
15:39<pika>hi again...
15:39<the_hydra>ah i see
15:39<the_hydra>what's up pika?
15:39<pika>Alright, I tried to apply a 2.6.18 patch from blaisorblade to a source - with a huge load of rejects
15:40<pika>'previously applied' ....
15:40<the_hydra>all of them?
15:40[~]dgraves was about to search for hot chick rehi on the web.
15:40<dgraves>and see what langauage she programmed in.
15:40<the_hydra>dgraves: good luck...
15:40<rasix>wew.. even i can't open
15:41<dgraves>permission denied, ENOUSER!
15:41<pika>rasix: Use 'blaisorblade' as a google search item, there's a mirror out there
15:41<rasix>www.blabla cannot be found
15:41<rasix>i see
15:41<jdike> is there
15:42<pika>the_hydra: I found the linode guest kernel 2.6.19 out there, so I guess it should be possible to apply a skas3 patch (blaisorblade?) onto the 2.6.19 source
15:42<pika>jdike: I get a timeout on this URL
15:42<the_hydra>pika: IMHO yes..
15:42<jdike>no problem here
15:42<the_hydra>jdike: we suffered problem
15:42<pika>If I try to do so I get a huge pile of rejects
15:43<the_hydra>due to earthquake in Taiwan days ago
15:43<the_hydra>pika: try to ask jeff
15:43<pika>I would love to wait for blaisorblades official 2.6.19 patch, but he wrote something of 'too busy' somewhere - that's a very vague time range? :-)
15:43<pika>jeff: how can I apply a lower uml skas patch on a 2.6.19 source?
15:44<rasix>thanks pika
15:45<jdike>pika, actually, I dunno
15:45<jdike>BB is the keeper of the skas patch
15:45<dgraves>jdike: when is skas52 coming out?
15:45<jdike>sometime in 2873
15:45<the_hydra>dgraves: when you finished Oracle Millenium edition ;)
15:45<jdike>set your alarms
15:45<dgraves>wouldn't that be like trillenium edition? :)
15:46<dgraves>jdike: what about skas4?
15:46<the_hydra>dgraves: oh right
15:46<dgraves>i want no speed loss, infinite debugging, and all the power of virtualization.
15:46<the_hydra>then dgrave the-300th will be the one who tested it ;)
15:46<jdike>After SMP and a KVM client
15:46<pika>I just want a skas patch without a sleep timer problem
15:47<dgraves>Oooo.... nice.
15:47<jdike>because then, I'll start worring about using the namespace stuff for UML
15:47<the_hydra>be happy allan, your descedants will see SKAS 52 ;)
15:47<dgraves>my descendents will finally see the ultimate uml to rule them all!
15:47<jdike>and that will require some extra support in the host
15:47<dgraves>yeah, ruling them all is tough.
15:47<the_hydra>dgraves: :)
15:48<dgraves>jdike: so what's on the map after that then? is that like 2008?
15:48<dgraves>will you ever consider uml done?
15:48<jdike>I just listed the major items on the list
15:48<jdike>after that, I dunno
15:48<the_hydra>jdike: any news from Bodo Stroeser about UML port into VT/Pacifica?
15:49<jdike>Bodo wasn't doing that, I don't think
15:49<jdike>and anyway, I have a VT port
15:49<the_hydra>oh wrong man I guess
15:50<the_hydra>with VT port, are you still staying with ptrace approach?
15:50<jdike>the whole point is to avoid ptace
15:51<the_hydra>switching to user mode guest, mean vm_enter?
15:51<the_hydra>*now means*?
15:51<jdike>and system calls to the host are done with VMCALL
15:51<the_hydra>so you patch the related IDT entries with call to vm_enter/exit?
15:52<the_hydra>vmcs definition?
15:52<jdike>The IDT entries point to UML entry points
15:52[~]the_hydra trying to recall what he had studied about VT
15:52<jdike>The VMCS is entirely separate
15:53<dgraves>jdike: will there be a fallback mechanism till everyone gets their host linux updated?
15:53<jdike>I think it will be a separate binary
15:53<the_hydra>jdike: ahh, IDT to UML entry points
15:53<jdike>it's not obvious to me that you can have it look for VT and then become a VT guest
15:54<the_hydra>jdike: but I am sure you now have a liberty to separate user and kernel mode into separate ring, no?
15:54<jdike>separate guest rings, yes
15:54<jdike>UML in guest ring 0
15:58|-|the_hydra__ [~a_mulyadi@] has joined #uml
15:58|-|the_hydra [~a_mulyadi@] has quit [Read error: Connection reset by peer]
15:58|-|the_hydra__ changed nick to the_hydra
16:01[~]dgraves thinks jdike is amazing.
16:01[~]the_hydra think dgraves is fabolous
16:03[~]dgraves waits for jdike.
16:03[~]jdike tries not to think
16:03<jdike>too hard
16:05<pika>About 2.6.19: Paolo / Blaisorblade just announced a skas patch for 2.6.19 today in the devel list
16:06<pika>cu guys
16:07<the_hydra>cya pika
16:07|-|pika [] has quit [Quit: [BX] We drink more beers than Norm on Cheers!]
16:56<rasix>the_hydra: ... still there buddy
17:01<the_hydra>yes I am here
17:01<the_hydra>what's up?
17:28|-|the_hydra [~a_mulyadi@] has quit []
17:28|-|rasix [~jeruk@] has quit [Quit: Leaving]
17:38|-|daxxar [] has joined #uml
17:38<daxxar>Hiya. :-)
17:38<daxxar>Is it possible to boot uml from a partition? (I.e. /dev/hda6)
17:38<daxxar>I tried root=/dev/hda6, but that barfed out :(
17:45<daxxar>Ah, I can mount /dev/hda6 on the host and boot from the dir
18:01<daxxar>Hmm, this isn't working properly. :|
18:02<daxxar>I'm trying to set up a debian-system inside UML, so I can later migrate it to the main system (so it's installed on it's own partition)
18:04<daxxar>I've used cdebootstrap on the partition, and got a halfway-working UML-system. This is the output:
18:04<daxxar>I'm curious why I get these:
18:04<daxxar>Failed to open 'root_fs', errno = 2; VFS: Mounted root (hostfs filesystem) readonly.
18:05<daxxar>And it fails on some modprobing, most likely because the system has no kernel installed in it's directories.
18:08<daxxar>And of course the xterm_open-thingies
18:14<richardw_>daxxar, use ubd0=/dev/hda6
18:15<daxxar>Then root=/dev/udb0?
18:16<richardw_>no. you don't need root
18:16<richardw_>uml uses ubd0 as root device
18:19<daxxar>Crap. I thought this was going to be easy, guess not.
18:19<daxxar>Seems my host server has changed gcc since last kernel compile, so I can't insert the new tuntap-module I compiled. :-P
18:19<daxxar>So I can't get networking working, which is why I wanted to use uml. :|
18:20<daxxar>FATAL: Error inserting tun (/lib/modules/2.6.12/kernel/drivers/net/tun.ko): Invalid module format
18:28<richardw_> /lib/modules/2.6.12/kernel/drivers/net/tun.ko has to be a module form the usermode kernel...
18:32<daxxar>No, I'm trying to modprobe it on the host, it doesn't have tuntap-support =(
18:35|-|richardw_ [] has quit [Quit: Leaving]
21:02|-|Newsome [] has joined #uml
21:04|-|ram [] has quit [Ping timeout: 480 seconds]
21:17|-|SiLLeNt [~FFdVv@] has joined #uml
21:34|-|SiLLeNt [~FFdVv@] has quit [Quit: /raw kill -9 *]
22:59|-|VS_ChanLog [] has left #uml [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #uml
23:09|-|jdike [~jdike@] has quit [Quit: Leaving]
---Logclosed Thu Jan 04 00:00:51 2007