Back to Home / #uml / 2007 / 09 / Prev Day | Next Day
#uml IRC Logs for 2007-09-24

---Logopened Mon Sep 24 00:00:55 2007
00:34|-|balbir [~balbir@] has quit [Ping timeout: 480 seconds]
01:05|-|balbir [~balbir@] has joined #uml
01:21|-|balbir [~balbir@] has quit [Read error: Connection reset by peer]
03:01|-|tyler [] has joined #uml
03:41|-|netdrake [netdrake@2001:5c0:904b:ffff::1] has quit [Read error: Connection reset by peer]
03:42|-|netdrake [~netdrake@2001:5c0:904b:ffff::1] has joined #uml
03:47|-|netdrake [~netdrake@2001:5c0:904b:ffff::1] has quit [Remote host closed the connection]
03:49|-|netdrake [] has joined #uml
04:06|-|netdrake [] has quit [Read error: Connection reset by peer]
04:08|-|netdrake [] has joined #uml
04:12|-|tyler [] has quit [Ping timeout: 480 seconds]
04:14|-|netdrake [] has quit [Remote host closed the connection]
04:19|-|netdrake [] has joined #uml
04:22|-|tyler [] has joined #uml
04:26|-|netdrake [] has quit [Remote host closed the connection]
04:27|-|netdrake [] has joined #uml
04:33|-|netdrake_ [netdrake@2001:5c0:904b:ffff::1] has joined #uml
04:34|-|netdrake [] has quit [Remote host closed the connection]
04:48|-|flinkdeldinky [] has joined #uml
04:49|-|flinkdeldinky [] has quit [Remote host closed the connection]
04:55|-|netdrake_ [netdrake@2001:5c0:904b:ffff::1] has quit [Remote host closed the connection]
04:55|-|netdrake [] has joined #uml
05:06|-|netdrake [] has quit [Remote host closed the connection]
05:06|-|netdrake [] has joined #uml
05:12|-|netdrake [] has quit [Remote host closed the connection]
05:20|-|netdrake [netdrake@2001:5c0:904b:ffff::1] has joined #uml
06:29|-|Baltam35 [] has quit [Remote host closed the connection]
06:45|-|Baltam [] has joined #uml
07:33|-|dang [] has quit [Quit: Leaving.]
07:36|-|kokoko1 [~Slacker@] has quit [Ping timeout: 480 seconds]
07:48|-|krau [~cktakahas@] has joined #uml
08:04|-|dang [] has joined #uml
08:50|-|Baltam [] has quit [Remote host closed the connection]
09:06|-|hfb [~hfb@] has quit [Quit: Leaving]
09:08|-|Baltam [] has joined #uml
09:47|-|kokoko1 [~Slacker@] has joined #uml
10:05|-|Groundy [] has joined #uml
10:05|-|Baltam [] has quit [Remote host closed the connection]
10:05|-|Groundy changed nick to Guest1004
10:06|-|Guest1004 changed nick to groundy
10:06|-|groundy changed nick to Groundy
10:06<Groundy>hello all,
10:06<Groundy>I recently got my 2.4 uml upgraded to 2.6
10:07<Groundy>except that now some programs (like aptitude and apache2) are segfaulting on startup
10:07<peterz>2.6 is rather ambiguous
10:08<Groundy>host is
10:08<Groundy>uml was 2.4.26-1um
10:08<Groundy>is now
10:08<Groundy>(and it looks like it was for a moment)
10:08<peterz>right, not the latest but sort-of recentish
10:09<Groundy>this is a hosted box, so I have full access to inside uml, but very little outside,
10:09<peterz>jeff might know, I just remember other people having similar issues
10:09<peterz>didn't pay too much attention, so I missed the solution, if any
10:09<Groundy>would it be in a chat log somewhere?
10:10<Groundy>or on one of the lists?
10:10<peterz>just sit around for a while, maybe someone else remembers/knows more
10:11<dang>Jeff hasn't shown up yet today...
10:11<peterz>yeah, I know
10:11<dang>When he does, you might ask him again.
10:12<Groundy>cool, most of the host stats I've only managed to find in dmesg in syslog
10:15<Groundy>aiiie, soft lockup detected on CPU
10:17|-|hfb [] has joined #uml
10:20|-|Baltam [] has joined #uml
10:23<Groundy>actually it seems to die after clone's
10:23<Groundy>if that possble?
10:24|-|SNy changed nick to SNy-test
10:25|-|SNy-test changed nick to SNy
10:32<Groundy>hrm, due to lack of tls support,
10:32<dang>Ah, makes sense.
10:33<Groundy>found this:
10:33<Groundy>works 100%
10:34<Groundy>okay, so what's the RIGHT way to fix it?
10:34<dang>Update your userspace inside your UML?
10:34<Groundy>is that it?
10:35<dang>Yeah. Your userspace is using the old 2.4 kernel TLS stuff, rather than the new 2.6 kernel hotness.
10:35<dang>(NPTL is the new API)
10:35<dang>Or, you can just leave the workaround in place...
10:36<Groundy>nahh, that's ugly, it's going to confuse debian's aptitude
10:36<dang>It may cause a bit of a performance issue, but I'm not sure...
10:36<dang>When I switched from 2.4 UML to 2.6 UML, I built a whole new userspace based on a 2.6 distro... That may or may not be a solution for you.
10:50|-|Baltam [] has quit [Ping timeout: 480 seconds]
10:55|-|SNy changed nick to SNy-test
10:55|-|SNy-test changed nick to SNy
11:05|-|SNy changed nick to SNy-test
11:05|-|SNy-test changed nick to SNy
11:12|-|jdike [] has joined #uml
11:12<jdike>Hi guys
11:13<peterz>hey jdike
11:13|-|SNy changed nick to SNy-test
11:13<peterz>jdike: Q, do you have lockdep support for um?
11:14|-|SNy-test changed nick to SNy
11:14<peterz>care to send it out?
11:14<jdike>let me make sure we're talking about the same thing
11:14<jdike>Ingo's lock ordering proof stuff?
11:15<jdike>that's there, just enable it
11:15<peterz>hmm, must be blind today
11:15<peterz>would not be the first elephant I missed
11:16<jdike>at least the bootup checking was enabled by default for a while and I saw all these self-tests go by
11:16<jdike>never actually really played with it
11:16<peterz> │ │ [*] Lock debugging: prove locking correctness │ │
11:16<peterz>don't get that on ARCH=um
11:17<jdike>OK, hold on
11:17<SNy>Uh, sorry for the disturbance, guys. I needed to fix a PEBKAC with NickServ here.
11:18<peterz>jdike: I would have thought you'd made extensive usage of it when doing the skas-smp work
11:19<jdike>the skas-smp work was just getting something booting
11:19<peterz>ah, the correctness part still has to come :-)
11:21|-|Groundy [] has left #uml []
11:22<jdike>OK, there is a maze of dependencies around LOCKDEP_SUPPORT
11:23<jdike>but it depends on TRACE_IRQFLAGS_SUPPORT, which isn't there
11:24<peterz>a well, I'll hack around it for the time being
11:31|-|Baltam [] has joined #uml
11:31|-|Baltam [] has quit [Remote host closed the connection]
11:44|-|kos_tom [] has joined #uml
11:48|-|Baltam [] has joined #uml
11:49|-|hfb [] has quit [Ping timeout: 480 seconds]
11:59|-|hfb [] has joined #uml
12:16|-|hfb [] has quit [Ping timeout: 480 seconds]
12:16|-|ram [] has joined #uml
12:25|-|hfb [] has joined #uml
12:33[~]dgraves would love to have locking correctness. :)
12:39|-|dang [] has quit [Remote host closed the connection]
12:40|-|tyler [] has quit [Ping timeout: 480 seconds]
12:50|-|dang [] has joined #uml
12:59|-|hfb [] has quit [Ping timeout: 480 seconds]
12:59|-|tyler [] has joined #uml
13:09|-|tyler [] has quit [Ping timeout: 480 seconds]
13:10|-|hfb [] has joined #uml
13:11|-|tyler [] has joined #uml
13:16|-|Baltam [] has quit [Remote host closed the connection]
13:21<kokoko1>hi hi
13:26|-|hfb [] has quit [Quit: Leaving]
13:26|-|hfb [] has joined #uml
13:31|-|Baltam [] has joined #uml
13:33|-|Baltam [] has quit [Remote host closed the connection]
13:41|-|Baltam [] has joined #uml
13:46|-|Baltam [] has quit [Remote host closed the connection]
13:54|-|Baltam [] has joined #uml
14:05<ctrace>< Groundy> found this:
14:05<ctrace>yup i ran into the same problem /w TLS
14:05<ctrace>expect and tclsh would just segfault upon trying to run them
14:06<ctrace>i found a couple of webpages documenting this fix as well:
14:06<ctrace>and also this one:
14:07<ctrace>increasing the amount of memory in a UML instance helps in some other cases too, like if you try to run tshark in a UML /w only 32MB it will just abort but 128MB works fine
14:07|-|tyler [] has quit [Ping timeout: 480 seconds]
14:10<jdike>UML has had TLS support for quite a while
14:25|-|tyler [] has joined #uml
15:11<ctrace>jdike: yup i noticed from dang's comments...seems like i am probably running a 2.6 kernel with old 2.4 userspace cruft
15:12<jdike>2.4 userspace shoud run
15:15<ctrace>hmm yes, it does seem to, but i did have to mv /tls / in order to run some apps (such as expect/tclsh)
15:16<jdike>what version of UML?
15:46|-|mgross [] has joined #uml
15:50<ctrace>jdike: 2.6.20
15:54<jdike>have you tried anything closer to current?
16:01<ctrace>not yet, but i can give that a try
16:02<ctrace>is there a good place to find more updated UML filesystems ?
16:03<ctrace>(prefer debian)
16:05<ctrace>so far we have been using the debian 3.1 root fs at as well as a custom root fs built with debootstrap
16:05|-|hfb [] has quit [Read error: Connection reset by peer]
16:07|-|hfb [] has joined #uml
16:09|-|dang [] has quit [Quit: Leaving.]
16:19<jdike>you could just upgrade it from that to whatever is more current
16:19<jdike>Debian makes it fairly easy
16:43|-|tyler [] has quit [Remote host closed the connection]
16:46|-|tyler [] has joined #uml
17:27|-|kos_tom [] has quit [Quit: I like core dumps]
17:56|-|tyler [] has quit [Remote host closed the connection]
18:20|-|mgross [] has quit [Quit: Leaving]
18:36|-|hfb [] has quit [Quit: Leaving]
19:38|-|Baltam [] has quit [Ping timeout: 480 seconds]
19:43|-|Baltam [] has joined #uml
19:55|-|dsoul [] has quit [Ping timeout: 480 seconds]
19:55|-|Baltam [] has quit [Ping timeout: 480 seconds]
20:17|-|Baltam [] has joined #uml
21:19|-|Baltam [] has quit [Ping timeout: 480 seconds]
21:50|-|Baltam [] has joined #uml
22:37|-|jdike [] has quit [Quit: Leaving]
22:58|-|VS_ChanLog [] has left #uml [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #uml
23:23|-|cgf [] has joined #uml
23:25|-|Electric1lf [] has quit [Ping timeout: 480 seconds]
23:45|-|ElectricElf [] has joined #uml
---Logclosed Tue Sep 25 00:00:37 2007