Back to Home / #uml / 2007 / 05 / Prev Day | Next Day
#uml IRC Logs for 2007-05-08

---Logopened Tue May 08 00:00:07 2007
---Daychanged Tue May 08 2007
00:00|-|da-x [] has joined #uml
04:08|-|kos_tom [] has joined #uml
04:09|-|polyonymous [] has quit [Ping timeout: 480 seconds]
04:20|-|polyonymous [] has joined #uml
06:10|-|krau [~cktakahas@] has joined #uml
07:16|-|kybe [] has quit [Remote host closed the connection]
07:25|-|mjf [] has joined #uml
07:26|-|kybe_ [] has joined #uml
08:11|-|jdike [] has joined #uml
08:11<jdike>Hi guys
08:20<dgraves>what's the first order of business today?
08:28<jdike>print up some more Chinese flash cards
08:28<jdike>then fix my irqstack locking
08:28<dgraves>hmm... sounds good to me.
08:28<dgraves>i'll get started on the flash cards.
08:29<jdike>glad you approve
08:29<dgraves>do you have a computer based flash program? i used to use one for japanese.
08:29<dgraves>it would pop them up every 15 min or so.
08:29<jdike>OO, xpdf, and lpr
08:29<jdike>physical cards
08:29<jdike>for when my laptop isn't convenient
08:30<dgraves>just curious. i mean, you don't work in front of a computer or anything. :)
08:30<jdike>airports and planes
08:30<dgraves>well, i mean for the times your laptop is convienent.
08:31<dgraves>otherwise its no good.
08:31<jdike>then I work
08:31<jdike>otherwise what's no good?
08:31<dgraves>couldn't do 2 flashcards every 15 min? :)
08:31<dgraves>computer based flash cards are no good when you don't have a computer.
08:31<dgraves>its uh.. kinda obvious. :)
08:32<jdike>PHYSICAL cards
08:32<jdike>on paper
08:33[~]dgraves thinks jdike is confused about what he's saying. :)
08:33<dgraves>but its okay. i was really just curious.
08:33<dgraves>anyway, your flash cards are at the printer now.
08:33<dgraves>you gonna come by and pick them up?
08:34<jdike>no they're not
08:34<dgraves>well, give me your cups info and i'll print it there. and your root password.
08:34<jdike>I just induced a paper jam and need to figure out how to start things up in the middle
08:34[~]dgraves figures this is okay, b\c its probably a uml box anyway. :)
08:34<dgraves>that's no good.
08:36<dgraves>how is the chinese coming anyway?
08:36<dgraves>any closer to figuring out where the bathroom is?
08:37<jdike>well, I can read the male/female indicators
08:37<jdike>that's important
08:37<jdike>and one test I failed in Tokyo once
08:54|-|hfb [] has joined #uml
08:54|-|hfb [] has quit [Remote host closed the connection]
09:09<dgraves>jdike: eep!
09:09<dgraves>the worst i did was insist in class that i wanted to be a woman. :-P
09:10[~]dgraves failed that test at college though.
09:10|-|rasix [~jeruk@] has joined #uml
09:10<jdike>that was roughly my reaction
09:12<dgraves>that was also the reaction of the japanese students in that class... :)
09:35|-|Groundy [] has joined #uml
09:36|-|Groundy [] has left #uml []
09:36|-|Groundy [] has joined #uml
09:36<Groundy>I'm having issues with arp inside uml,
09:37<Groundy>and I think I found a solution, but wanted to double check
10:14<peterz>jdike: where are UM_PROT_READ and co supposed to come from?
10:15<jdike>compilation problems?
10:15<peterz>awkay, /me goes fix that
10:15<peterz>as of a few hours ago
10:15|-|Groundy [] has left #uml []
10:16<jdike>2.6.21 builds for me
10:16<peterz>ok, I'll repull
10:16<peterz>just my luck to get an unlucky split
10:16<peterz>jdike: thanks for sending that um scsi thing along
10:16<jdike>I shouldn't have sent in an inconsistent patch set
10:18<peterz>nope, still; maybe toss my build dir
10:19<peterz>didn't help either
10:19<peterz>awell, I'll go edit that file
10:19<jdike>I don't see my latest stuff on gitweb
10:19<peterz>true; your latest series hasn't made it yet
10:20<jdike>AFAICS, I introduce definitions of those in the same patch as their first use
10:21<peterz>um/kernel/tlb.c is not wanting to compile; I'll go look at patches
10:22<jdike>that's where I think I introduced and defined them
10:23<peterz>that does seem to suggest so
10:23<jdike>and this bit is what's not compiling?
10:23<jdike>+ prot = ((r ? UM_PROT_READ : 0) | (w ? UM_PROT_WRITE : 0) |
10:23<jdike>+ (x ? UM_PROT_EXEC : 0));
10:23<jdike>so those aren't in arch/um/include/user-constants.h?
10:25<jdike>remove it and rebuild
10:25<jdike>maybe this is a build bug
10:25<peterz>I rm -rf'ed my whole build dir; but I can try this one file
10:25<jdike>well, that should have done it
10:26<peterz>aahh, its in O=. but not in O=uml-build
10:26[~]peterz removes both
10:27<jdike>have you done a non-O= build in this tree?
10:27<peterz>nope; that screws over the O=someotherdir builds
10:27<peterz>forces you to do a mrproper
10:27<jdike>did you just say you had a $TOPDIR/arch/um/include/user-offsets.h?
10:28<jdike>OK, then you did, and mrproper didn't get rid of it
10:28<peterz>maybe add that to mrproper ?
10:28<jdike>I think so
10:29<peterz>build is happily chugging along
10:50|-|scoates [] has joined #uml
10:50<scoates>hello again
10:51<scoates>I have UML running.. however, when I try to log in at the console, when I type root<enter> I immediately get "Login timed out after 60 seconds." with no time to type the password. ideas?
10:55<scoates>erm.. stupid question: how do I check?
10:56<scoates> uml@adnagaporp:~$ linux.uml --version
10:56<scoates>it's from debian
10:56<jdike>can you see if 2.6.21 has the same problem?
10:58<scoates>this was working yesterday. I changed the networking options and rebooted, and it broke... )-:
10:58<scoates>I'll try to find 2.6.21
10:58<jdike>does it happen on every boot?
10:59<jdike>today, but not yesterday
10:59<scoates>should I mount the ubd and undo my network changes?
10:59<scoates>(although I'm pretty sure they're correct)
11:00<jdike>do you try to log in immediately after it boots?
11:00<scoates>and I tried to log into one that had been booted for ~16h
11:00<scoates>same behaviour
11:01<jdike>have you changed the host system clock since yesterday?
11:01<scoates>fwiw, my network settings actually.. work.. I can ping the uml
11:01<scoates>no, I didn't change it
11:01<scoates>I didn't remember to set up ssh, though
11:01<scoates>so console is my only option atm
11:01<jdike>might ntp have adjusted the time more than 1 minute?
11:01<scoates>unlikely, but possible I guess
11:02<jdike>I have seen something like this, but it got worse over time
11:02<scoates>similar to that
11:02<jdike>i.e. you could log in immediately on boot, but wait a bit, and you started getting the 60-sec timeouts
11:02<scoates>... I changed the hostname of the uml before reboot
11:03<scoates>I installed vim and a few other packages.. might have installed ntp, but 100% sure
11:04<scoates>line_ioctl: tty0: unknown ioctl: 0x541e
11:04<scoates>line_ioctl: tty0: unknown ioctl: 0x5603
11:04<scoates>hwclock is unable to get I/O port access: the iopl(3) call failed.
11:05<scoates>that might be relevant
11:05<jdike>linux.uml --showconfig | grep REAL_TIME
11:05<jdike>it's not
11:05<scoates>uml@adnagaporp:~$ linux.uml --showconfig | grep REAL_TIME
11:06<jdike>OK, what happens if you undo the networking changes?
11:06<scoates>let's find out (-:
11:10<scoates>booting now
11:11<scoates>undid my network changes and now I can log in
11:12<scoates>however, if I start the network now, the network is fine
11:12<scoates>(I'm sure.. haven't actually tried it yet, but network was fine yesterday)
11:12<jdike>run time sleep 10
11:12<scoates>real 0m10.057s
11:12<scoates>user 0m0.000s
11:12<scoates>sys 0m0.000s
11:13<jdike>does date tell you the same thing as the host?
11:13<scoates>timezones are different, but otherwise yes
11:14<jdike>is there an FC equivalent of that file?
11:15<scoates>equivalent of what file?
11:15<scoates>this is a debian host with a ubuntu guest
11:16<scoates>ah.. no idea.. I haven't used redhat in years
11:16<scoates>there's like /etc/system/network, maybe?
11:17<jdike>FWIW, I did the equivalent AFAICS, and have no trouble
11:17<jdike>can you put that filesystem someplace where I can grab it?
11:17<scoates>with or without the net changes?
11:18<jdike>comment them out so I know what they are
11:18<scoates>ah.. already done (-:
11:18<scoates>it's a gig.. let me upload it somewhere sane
11:20<scoates>think this might have to do with not having a hostname set?
11:20<scoates>(guessing here, obviously)
11:20<jdike>no idea
11:20<jdike>this is just bizarre
11:20<scoates>I agree (-:
11:21<jdike>in the meantime can you check 2.6.21?
11:21<scoates>I'll try to find a deb package for it..
11:25|-|rasix [~jeruk@] has quit [Quit: Leaving]
11:27[~]scoates curses the slow disk in this machine
11:28<peterz>I doubt debian has packages for anything less than a year old
11:28<scoates>looking at backports, now
11:28[~]peterz runs
11:30<scoates>no package.. )-:
11:30<jdike>it's easy enough to build from source
11:37[~]jdike starts a kernel build loop and disappears for a bit
11:37<scoates>I'm bzipping the disk image anyway
11:37<scoates>PIII 700 + bzip = a long time
11:38|-|scoates_ [] has joined #uml
11:40|-|scoates [] has quit [Read error: Operation timed out]
12:03<scoates_>jdike: see corner
12:12|-|kos_tom_ [] has joined #uml
12:36<jdike>got it
12:39<jdike>what's the root passwd?
12:39<jdike>pm it if you want
12:40<jdike>So, uncomment the whole iface eth0 bit?
12:41<scoates_>yeah, and you'll have to comment the line that has dhcp in it
12:42<scoates_>ok.. a bit of new info:
12:42<scoates_>if I do that within the UML and /etc/init.d/networking restart
12:42<scoates_>net works fine
12:42<scoates_>but if I reboot (within UML), the console won't let me log in
12:42<scoates_>timeout error as before
12:43<scoates_>for anything I enter into the "login:" prompt (not just root)
12:53<scoates_>... and I do have my hostname set, now
13:04<jdike>no problem here
13:04<jdike>have you checked 2.6.21?
13:04<scoates_>no.. downloading new kernel.. the mirror's slow
13:05|-|scoates_ changed nick to scoates
13:21<scoates>fwiw, it only happens if I automatically start the interface. If I don't eth0 auto, and then ifup eth0 when I log in, it works fine.
13:28|-|max [] has joined #uml
13:32<max>Hi everyone
13:33<max>I need help to setup a connexion in my uml
13:33<max>can anyone help me ?
13:34<jdike>what's the question
13:35<max>well I tape the following command Noyaux/linux-2.4.27 umid=ordi1 ubd0=cow02,SysFich/root_fs_slack8.1 eth0=tuntap,tap0
13:35<max>to run my uml but when i configure eht0 in uml
13:35<max>i can't ping the tap0 interface
13:36<jdike>have you already set up the tap0 device on the host?
13:36<max>yes with tunctl
13:36<max>and then with ifconfig tap0 x.X.X.X
13:37<jdike>the ifconfig succeeds?
13:37<jdike>and nothing in dmesg?
13:37<jdike>the UML ifconfig
13:37<max>oh i didn't check in dmesg
13:38<max>the uml ifconfig succeed
13:38<jdike>do you have a route to the UML through the tap device?
13:38<max>but indmesg i got a msg with shorewall which doesn't allow the request from uml
13:38<jdike>on the host
13:39<max>oki i will desactivate shorewall ^^
13:39<jdike>yeah, you need to poke a hole in your firewall for the UML
13:39<max>oki thanks
13:39<max>and sorry if my english isn't really good it's not my first language
13:40<jdike>it's better than my grasp of whatever your first language is
13:41<max>french ^^
13:41<jdike>there's my grasp of French
13:41<jdike>plus assorted other words
13:42<max>it's a start :p
13:42<jdike>not a good one
13:43<jdike>I can actually get around passibly in France
13:43<jdike>based on English, Spanish, and smatterings of other languages
13:43<max>oh nice
13:47<max>oki here is just what i have done on the host : tunctl -d tap0 then tunctl -u 0 then ifconfig tap0 up then Noyaux/linux-2.4.27 umid=ordi1 ubd0=cow02,SysFich/root_fs_slack8.1 eth0=tuntap,tap0
13:47<max>and in uml ifconfig eth0 up
13:47<max>but still i can't ping
13:47<max>and nothing new in dmesg
13:48<jdike>routes in both directions?
13:48<max>what do you mean ?
13:48<jdike>a route on the host to the UML through tap0
13:49<jdike>and a route in UML to the outside world through eth0
13:49<max>ah i didn't do that
13:49<max>is it a command line with the route commande
13:49<jdike>route add -host dev tap0
13:50<jdike>on the host
13:52<max>and do i have to add a route on uml either ?
13:52<jdike>don't think so
13:52<max>well the route is added
13:52<jdike>it would have complained about host unreachable
13:53<max> * UH 0 0 0 tap0 --> first line when i type route
13:53<max>by the way route is the french word for road ^^
13:55<jdike>english too
13:55<jdike>that looks OK
13:55<max>oh i didn't know
13:55<jdike>not as general though
13:56<jdike>so, no pings still?
13:56<jdike>tcpdump tap0
13:56<jdike>any other firewall, BTW?
13:57<max>no no other firewall ^^
13:57<jdike>iptables -L says nothing?
13:58<max>ok i think i will have to apologize
13:58<max>iptable was started
13:58<max>sorry my mistake :-(
13:58<max>but now it pings
13:59<max>i'm on a mandriva i installed it yesterday and i didn't know that two firewalls were on
14:00<max>aniway thanks a lot for your help
14:08|-|max [] has quit [Remote host closed the connection]
15:17[~]jdike spots a very basic bug
15:17<jdike>and wonders why it was so hard to make it break
15:46|-|mjf [] has quit [Quit: leaving]
15:48|-|kos_tom_ [] has quit [Quit: I like core dumps]
16:00|-|max [] has joined #uml
16:00|-|max [] has quit []
16:02|-|max [] has joined #uml
16:02<max>Hi again everyone
16:03<max>can anyone help me with my uml which doesn't want to ping the gateway of my own lan and of course
16:13<jdike>have you set up ip forwarding and proxy arp?
16:23|-|ram [] has joined #uml
16:23|-|max [] has quit [Ping timeout: 480 seconds]
16:23|-|maxlupin [] has joined #uml
16:24<maxlupin>sorry jdike a mistake with my access point
16:25<maxlupin>yes i activate ip_forwarding and arp with these commands : echo 1 > /proc/sys/net/ipv4/conf/tap0/proxy_arp & arp -Ds eth0 pub
16:25<jdike>how about IP forwarding?
16:26<maxlupin>is it the option in my host kernel that i have to activate
16:27<jdike>/proc/sys/net/ipv4/ip_forward needs a 1
16:27<maxlupin>oh yes i forgot this line yes i did
16:28<jdike>OK, you can ping the host?
16:29<maxlupin>i can ping eth0 of my host
16:29<maxlupin>and from my host i can ping eth0 of my uml
16:29<jdike>tcpdump the host's eth0, ping the LAN gatewat, and see if the UML packets reach the host eth0
16:31<maxlupin>yes it reachs eth0 on my host when i ping my gateway from my uml
16:31<maxlupin>IP > ICMP echo request, id 19456, seq 0, length 64
16:32<maxlupin>IP > 61067+ PTR? (42)
16:32<maxlupin>IP > 61067 NXDomain 0/1/0 (119)
16:32<maxlupin>IP > 1661+ PTR? (42)
16:32<maxlupin>IP > 1661 NXDomain 0/1/0 (119)
16:32<maxlupin>IP > 2576+ PTR? (43)
16:32<maxlupin>IP > 2576 NXDomain 0/1/0 (120)
16:33<jdike>OK enough
16:33<maxlupin>sorry for the flood
16:33<jdike>go to the gateway and see if those packets are getting there
16:34<jdike>who is 1.10?
16:34<maxlupin>it's my eth0 on the host
16:35<jdike>they're on different subnets
16:35<maxlupin>yes is it a problème ?
16:35<jdike>the rest of the network will need a route to it
16:35<maxlupin>oh ok
16:35<jdike>everyone else needs to know that the host is the gateway to 192.168.2.x
16:36<maxlupin>so i have to configure a route on my gateway
16:36<jdike>any local system that you want to talk to
16:38|-|scoates_ [] has joined #uml
16:42<maxlupin>oki it works now thanks a lot again jdike
16:43<maxlupin>another question completely different
16:44<maxlupin>is it possible to squeeze the login procedure at startup of uml ?
16:44|-|scoates [] has quit [Ping timeout: 480 seconds]
16:45|-|scoates_ changed nick to scoates
16:46<scoates>jdike: well, I got compiled.. checking the boot w/network now
16:47<maxlupin>i mean that i don't want to write root for the login procedure i just want to be in a shell with root privileges
16:47<maxlupin>at startup
16:48<scoates>jdike: and seems to fix it.. whatever "it" was
16:48<scoates>the problem was affecting ssh, too
16:48<scoates>getting timeouts
16:48<jdike>be nice to know what the problem was, but as long as it's gone, I guess I'm happy
16:49<scoates>heh.. me too (-:
16:49<jdike>maxlupin, nothing to do with UML
16:49<jdike>but if you changed getty to bash in inittab, that might work
16:50<scoates>could also boot into single user mode somehow, no?
16:53<jdike>sure, but I assumed you wanted the system fully booted
16:53<jdike>init=/bin/bash will give you a root shell
16:55<scoates>well, not me (-:
16:56<scoates>I meant to "squeeze" the login
16:56<jdike>then just have inittab run bash instead of getty
16:57|-|maxlupin [] has quit [Ping timeout: 480 seconds]
17:16|-|kos_tom [] has quit [Quit: I like core dumps]
17:38|-|jdike [] has quit [Quit: Leaving]
18:36<scoates>any idea why I wouldn't have a loopback device?
18:41<scoates>typo in the ubuntu image (-:
19:05|-|besonen_mobile [] has quit [Quit: Leaving]
19:25|-|Nem^ [] has joined #uml
19:28|-|scoates [] has quit [Quit: scoates]
19:32|-|scoates [] has joined #uml
20:51|-|albertito [] has quit [Ping timeout: 480 seconds]
20:56|-|albertito [] has joined #uml
21:08|-|Nem^ [] has quit [Ping timeout: 480 seconds]
21:16|-|Nem^ [] has joined #uml
21:44|-|scoates [] has quit [Ping timeout: 480 seconds]
---Logclosed Tue May 08 22:20:28 2007
---Logopened Tue May 08 23:25:41 2007
23:25|-|mikegrb [] has joined #uml
23:25|-|Ekipa kanalu #uml: Wszystkich: 15 |-| +op [0] |-| +voice [0] |-| normalnych [15]
23:25|-|Kanal #uml zsynchronizowany w 1 sekundy
23:27|-|ElectricElf [] has quit [Ping timeout: 480 seconds]
23:32|-|ElectricElf [] has joined #uml
---Logclosed Wed May 09 00:00:06 2007