Back to Home / #uml / 2007 / 05 / Prev Day | Next Day
#uml IRC Logs for 2007-05-04

---Logopened Fri May 04 00:00:18 2007
01:48|-|Internat [biteme@c210-49-250-210.ipswc1.qld.optusnet.com.au] has quit [Read error: Connection reset by peer]
01:49|-|Internat [biteme2@c210-49-250-210.ipswc1.qld.optusnet.com.au] has joined #uml
03:22|-|da-x [karrde@bzq-82-81-114-146.red.bezeqint.net] has quit [Remote host closed the connection]
03:22|-|da-x [karrde@bzq-82-81-114-146.red.bezeqint.net] has joined #uml
04:08|-|motp [~motp@83.236.181.13] has joined #uml
04:09|-|polyonymous [~hacker@pd953907d.dip0.t-ipconnect.de] has quit [Ping timeout: 480 seconds]
04:20|-|polyonymous [~hacker@pD953B732.dip0.t-ipconnect.de] has joined #uml
04:28|-|Ancalagon [~PtitKosmi@215.208-78-194.adsl-fix.skynet.be] has joined #uml
04:51|-|kokoko1 [~Slacker@203.148.65.8] has joined #uml
05:00|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml
05:23|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has quit [Quit: aroscha]
05:27|-|rasix [~jeruk@167.205.33.253] has joined #uml
05:37|-|kokoko1 [~Slacker@203.148.65.8] has quit [Ping timeout: 480 seconds]
05:51|-|poppyer [~user@clpc78.comlab.ox.ac.uk] has joined #uml
05:52<poppyer>hi, all , i am frequently getting "[Errno 12] Cannot allocate memory", any idea?
06:01|-|kokoko1 [~Slacker@203.148.65.8] has joined #uml
08:20|-|aroscha [~aroscha@213.235.234.114] has joined #uml
08:25|-|jdike [~jdike@pool-71-174-247-179.bstnma.fios.verizon.net] has joined #uml
08:25<jdike>Hi guys
08:25<jdike>peterz, there?
08:25<peterz>yep
08:25<jdike>those two UML patches
08:25<jdike>didn't I see those a long time ago?
08:26<peterz>yep
08:26<peterz>still not merged
08:26<jdike>hmmm
08:26<peterz>I've been carrying them in my networked swap series
08:26<peterz>and that work is so controversial that nobody comments on it
08:26<peterz>I'm being dug a silent grave
08:26<peterz>or something
08:27<jdike>well, there's no explicit merge request, so maybe that's why
08:27<jdike>I'll stick them in my series and push them in
08:27<peterz>yeah, I put that in this one
08:27<peterz>something like; if no maintainer comments, I'll see that as acceptance and push for merge
08:28<jdike>you sent to a bunch of lists, but not to akpm or linus
08:28<peterz>akpm was included
08:28<jdike>oh yeah
08:28<jdike>missed that
08:28<peterz>Linus thinks I'm daft
08:28<peterz>well, he thinks the concept is
08:32<jdike>OK, I'll drop them in my tree and see if they end up in -mm
08:32<jdike>if not, then I'll send them myself
08:32<caker>jdike: hello
08:32<peterz>ok, thanks
08:32<jdike>caker, how's about that boot hang?
08:33<caker>jdike: still nothing, but I've got something else ...
08:33<caker>http://www.linode.com/forums/viewtopic.php?t=2729 <-- doesn't happen with 2.4-um
08:34<jdike>hmm
08:35<jdike>let me see if I see that here
08:38<jdike>there's absolutely nothing running
08:38<jdike>except the ps
08:39<jdike>nothing has accumulated any time
08:42<caker>I booted Ubuntu with init=/bin/bash and its loadavg is 0
08:42<caker>I've been getting a lot of reports of this; so perhaps it's something with one of the popular distros
08:43<jdike>I'm surprised people notice
08:43<caker>people freak...
08:44<caker>ZOMG LOADAVG
08:44<jdike>the OP is running Fedora
08:44<jdike>hehe
08:44<jdike>and I've got FC6 looking like it's settling down to 0
08:44<caker>what could it be [in userspace] that doesn't do this under 2.4?
08:44<caker>something TLS/NPTL related?
08:44<jdike>something would be getting time
08:45<jdike>unless it's a short-lived process
08:45<jdike>but it would have to be there ~40% of the time
08:45<jdike>and pids would be increasing
08:46<jdike>root 2548 0.0 0.1 1620 424 tty0 Ss+ 22:11 0:00 /sbin/mingetty tty0
08:46<jdike>root 2629 0.0 0.9 8024 2380 ? Ss 22:20 0:00 sshd: npk [priv]
08:46<jdike>81 processes ran and exited between that getty and the sshd
08:48[~]caker tries with FC6
08:48<jdike> 09:44:03 up 8 min, 1 user, load average: 0.48, 0.26, 0.12
08:48<jdike>hmm
08:48<jdike>not tending towards 0
08:54<jdike>gotta be a kernel thing
08:54<jdike>I just watched top for a few minutes - no random processes coming and going
08:55<jdike>load ave ~.4 - .5
08:55<caker>it seemed fine with my default FC6 depoyment .. going to pull in updates to see if that triggers it
08:57<poppyer>hi, is there any way to open a port forwarding when using eth0:slirp?
09:00|-|nessie [~nessie@lucifer.nerdfest.org] has quit [Ping timeout: 480 seconds]
09:11<caker>]# uptime
09:11<caker> 10:11:22 up 9 min, 1 user, load average: 0.30, 0.29, 0.19
09:12[~]caker stops everything
09:16<caker>http://rafb.net/p/nbSdBN63.html <-- jdike
09:16<jdike>yeah
09:16<jdike>nothing there except the load
09:20|-|Ancalagon [~PtitKosmi@215.208-78-194.adsl-fix.skynet.be] has quit [Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.3/2007031001]]
09:24<jdike>looks like kjournald
09:24<jdike>but top never shows it as D
10:09<jdike>the load settles down for ~5 minutes, then it picks up to ~.4
10:13|-|rasix [~jeruk@167.205.33.253] has quit [Quit: Leaving]
10:13<caker>and yet, no time is accounted for
10:14<caker>no phantom processes consuming pids, over the last 50 minutes, either
10:14<caker>root 2692 0.0 0.6 4208 932 tty0 R+ 10:15 0:00 ps aux
10:14<caker> 10:15:58 up 14 min, 1 user, load average: 0.38, 0.31, 0.20
10:14<caker>...
10:14<caker>root 2695 0.0 0.6 4212 932 tty0 R+ 11:12 0:00 ps aux
10:14<caker> 11:12:35 up 1:10, 1 user, load average: 0.38, 0.31, 0.27
10:15<caker>everything reports 0 cputime (killdashnine, no more cpu time)
10:15|-|aroscha [~aroscha@213.235.234.114] has quit [Quit: aroscha]
10:15<caker>my flow is like reentrand and preemptive multitasking
10:15<caker>*reentrant
10:20<jdike>I'm seeing something start being running every other clock tick
10:20<jdike>doesn't show in ps or top
10:24|-|hfb [~hfb@pool-72-67-156-130.lsanca.dsl-w.verizon.net] has joined #uml
10:24|-|hfb [~hfb@pool-72-67-156-130.lsanca.dsl-w.verizon.net] has quit [Read error: Connection reset by peer]
10:28<jdike>it's the idle thread
10:28<caker>huh
10:28<caker>kind of humorous, isn't it?
10:28<jdike>yeah :-)
10:29<jdike>40% idle, 60% really idle
10:31<jdike>man
10:32<jdike>innocent choices come back and bite you
10:32<jdike> idle_sleep(10);
10:32<jdike>meaning the idle loop sleeps for 10 secs, sees nothing going on, goes back to sleep
10:33<jdike>#define LOAD_FREQ (5*HZ) /* 5 sec intervals */
10:33<jdike>the load is averaged every 5 secs
10:35<caker>so .. LOAD_FREQ and idle_sleep should be the same, so load calc doesn't "see" the idle loop waking up?
10:35<jdike>no
10:35<jdike>they fall into sync with each other
10:36<jdike>so every other load calculation, it sees the idle loop in the middle of waking up and going back to sleep
10:36<jdike>hence a load of ~.5
10:36<jdike>resonance
10:37<caker>right
10:42<jdike>no
10:42<jdike>something else is going on
10:45|-|poppyer [~user@clpc78.comlab.ox.ac.uk] has left #uml [ERC Version 5.2 (IRC client for Emacs)]
10:47<jdike>change it to sleep for 61 seconds and it still wakes up every 10
10:55|-|motp [~motp@83.236.181.13] has quit [Quit: Leaving]
11:14|-|Internat-bed [biteme2@c210-49-250-210.ipswc1.qld.optusnet.com.au] has joined #uml
11:19|-|Internat [biteme2@c210-49-250-210.ipswc1.qld.optusnet.com.au] has quit [Ping timeout: 480 seconds]
11:23|-|Internat [biteme2@c210-49-250-210.ipswc1.qld.optusnet.com.au] has joined #uml
11:26|-|Internat-bed [biteme2@c210-49-250-210.ipswc1.qld.optusnet.com.au] has quit [Ping timeout: 480 seconds]
11:27|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml
12:10<jdike>caker, I may have a patch for you
12:10<caker>ok
12:10<caker>may? :)
12:11[~]caker fires up pizzahut.com for jdike
12:11<jdike>hehe
12:11<jdike>waiting for it to chill a bit longer before declaring victory
12:11<caker>what's the word?
12:12<jdike> 13:07:19 up 8 min, 1 user, load average: 0.00, 0.03, 0.02
12:15<jdike> 13:10:20 up 11 min, 1 user, load average: 0.00, 0.01, 0.00
12:15<jdike>I'll consider this to be goodness
12:15<jdike>let me remove all the debugging crap from the patch
12:17|-|ram [~ram@pool-71-117-233-61.ptldor.fios.verizon.net] has joined #uml
12:22<jdike>http://rafb.net/p/7i21LO86.txt
12:30<jdike>caker, you want the version I'm going to send in?
12:30<jdike>identical code, just changelogged and signed-off-by
12:34|-|ram [~ram@pool-71-117-233-61.ptldor.fios.verizon.net] has quit [Ping timeout: 480 seconds]
12:45<jdike>it takes 12 minutes, but loads do end up at 0
12:45<jdike> 13:34:18 up 12 min, 1 user, load average: 0.00, 0.00, 0.00
13:05|-|ram [~ram@bi01p1.co.us.ibm.com] has joined #uml
14:42|-|kalpak [~kalpak@203.145.159.44] has joined #uml
14:48|-|kalpak [~kalpak@203.145.159.44] has quit [Quit: Exiting]
14:55|-|mjf [~mjf@r5bb59.net.upc.cz] has joined #uml
15:40|-|mjf [~mjf@r5bb59.net.upc.cz] has quit [Quit: leaving]
15:44|-|kos_tom [~thomas@humanoidz.org] has joined #uml
15:59<caker>jdike: hey -- was afk
15:59[~]caker tries to grok the patch
16:15<caker>jdike: hmm, did your 3/3 patch make it out of your MUA?
16:17<jdike>yeah
16:17<jdike>I've got it in my uml folder
16:18<caker>it's MIA on my end, both from uml-user and LKML
16:18<caker>huh...
16:18[~]caker gives the evileye to Thunderbird
16:18|-|krau [~cktakahas@200.184.118.132] has quit [Ping timeout: 482 seconds]
16:19<caker>gd, sf.net is slow
16:19<jdike>it ain't there either
16:20<caker>http://sourceforge.net/mailarchive/forum.php?forum_name=user-mode-linux-devel&max_rows=25&style=ultimate&viewmonth=200705 <- it is there
16:20<jdike>Oh
16:20<jdike>OK
16:20<jdike>http://rafb.net/p/7i21LO86.txt
16:20<jdike>as well, plain text, easier to pipe into patch
16:21<caker>500 - Internal Server Error
16:21<caker>thanks sf.net
16:21<jdike>YOU ARE NOT MEANT TO HAVE THE PATCH
16:22<jdike></deep-voice-from-above>
16:24<caker>so, let me guess what the problem was (or make something up): it updates UML's ticker before doing the process accounting?
16:24<caker>s/problem was/fix does
16:25<jdike>I'm not actually sure
16:26<jdike>but there was something bad about allowing a schedule to happen in the timer interrupt before the tick was actually delivered
16:31|-|krau [~cktakahas@200.184.118.132] has joined #uml
17:32|-|krau [~cktakahas@200.184.118.132] has quit [Remote host closed the connection]
17:33|-|amitg [~gud@n119.user.cis.ksu.edu] has joined #uml
17:34<amitg>strange, kernel in UML won't boot if I give SLAB_DEBUG_* flag options to kmem_cache_create
17:34<amitg>if I remove those, it does.
17:34<amitg>is it a known issue? 2.6.18
17:36|-|karrde_ [karrde@bzq-88-152-109-61.red.bezeqint.net] has joined #uml
17:39|-|da-x [karrde@bzq-82-81-114-146.red.bezeqint.net] has quit [Ping timeout: 480 seconds]
17:47<jdike>nope
17:47<jdike>does it happen with 2.6.21?
17:47<amitg>not sure, but would be easy to check
17:47<amitg>add those flags to any kmem_cache_create
17:48<amitg>since those flags are not used in any code so far, it may not be tested yet
18:07|-|jdike [~jdike@pool-71-174-247-179.bstnma.fios.verizon.net] has quit [Quit: Leaving]
18:41|-|kos_tom [~thomas@humanoidz.org] has quit [Quit: I like core dumps]
19:52|-|ram [~ram@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds]
20:03|-|apic [~andi@paranerd-geekophren.name] has quit [Quit: Auszeit.]
20:46|-|Internat-bed [biteme2@c210-49-250-210.ipswc1.qld.optusnet.com.au] has joined #uml
20:46|-|Internat [biteme2@c210-49-250-210.ipswc1.qld.optusnet.com.au] has quit [Read error: Connection reset by peer]
22:59|-|VS_ChanLog [~stats@ns.theshore.net] has left #uml [Rotating Logs]
22:59|-|VS_ChanLog [~stats@ns.theshore.net] has joined #uml
---Logclosed Sat May 05 00:00:50 2007