Back to Home / #uml / 2007 / 08 / Prev Day | Next Day
#uml IRC Logs for 2007-08-01

---Logopened Wed Aug 01 00:00:42 2007
00:04|-|Internat85 [] has quit [Read error: Connection reset by peer]
00:04|-|Internat85 [] has joined #uml
03:01|-|horst [] has joined #uml
03:25|-|horst [] has quit [Remote host closed the connection]
03:57|-|baroni [] has quit [Ping timeout: 480 seconds]
04:46|-|HuK0B [~HuK0B@] has joined #uml
04:53|-|horst [] has joined #uml
04:56|-|horst [] has quit [Remote host closed the connection]
05:43|-|tyler [] has joined #uml
06:44|-|krau [~cktakahas@] has joined #uml
08:00|-|tyler [] has quit [Remote host closed the connection]
08:01|-|dgraves [] has joined #uml
08:24|-|dang [] has joined #uml
08:30|-|dang [] has quit [Quit: Leaving.]
08:51|-|kokoko1 [~Slacker@] has joined #uml
08:51<kokoko1>hi guys
09:00[~]kokoko1 updating uml kernels
09:01<kokoko1>we were running like more then 1.5 years :-s
09:01<kokoko1>going for on UML
09:01|-|dang [~dang@] has joined #uml
09:02<dgraves>kokoko1: yeah, we're on 2.6.14 here. :)
09:03<kokoko1>heh, successfully rebooted 6 uml, 2 hosts to go
09:04<kokoko1>with 3 , 5 UMLs respectively.
09:04<kokoko1>jdike said new kernel include some new uml code and lota fixes which will give some performance
09:05<kokoko1>dgraves, all of these are production UMLs :)
09:05<kokoko1>running client dedicated hosted, shared hosting, backups machines, mail server, ldap server, jabber blah blha :-s
09:07<kokoko1>for me as syadmin uml and custom kernel is way to cool, we also running xen on few hosts adn maintian xen for which we are using fedora folks kernel-xen it PITA
09:08<kokoko1>every day there is kernel-xen udpate and often it break and then we have fallback to last working kernel-xen :)
09:08<kokoko1>fedora folks rushes too much which brings more problems.
09:10<kokoko1>custom and self compiled kernel is blessing, no need to update if you do not feeling
09:17<dgraves>kokoko1: yeah, you're doing it right. ;)
09:17<dgraves>we're using htem for development.
09:24|-|balbir [~balbir@] has quit [Read error: Connection reset by peer]
09:27|-|tyler [] has joined #uml
09:43|-|balbir [~balbir@] has joined #uml
10:03|-|ram [] has joined #uml
10:56|-|da-x [] has quit [Ping timeout: 480 seconds]
10:58|-|jdike [] has joined #uml
10:59<jdike>Hi guys
11:06|-|da-x [] has joined #uml
11:10<jdike>back to writing LCEU paper
11:11<jdike>writing is a drag
11:56<dgraves>jdike: but you're good at it.
12:02[~]dgraves "enjoyed" the uml book..
12:03<jdike>in quotes?
12:09<dgraves>well, it was enjoyable, but it wasn't like a fiction book.
12:09<dgraves>very well written, readable, but not like a work of pleasure.
12:11<jdike>maybe that's why I'm not rich yet
12:11<jdike>I need to add some car chases and murders in the next edition, I guess
12:15|-|horst [] has joined #uml
12:15<dgraves>Ooo... there you go.
12:15<dgraves>did you make much off the book?
12:15<jdike>I ain't rich yet
12:15<jdike>is it possible to pass html through latex?
12:16<jdike>so that a latex -> html converter makes links correctly?
12:27<kokoko1>Checking that ptrace can change system call numbers...OK
12:27<kokoko1>Checking syscall emulation patch for ptrace...OK
12:27<kokoko1>Checking advanced syscall emulation patch for ptrace...OK
12:27<kokoko1>This host running (fc6) all of the uml now runing (6 of them)
12:28<kokoko1>i'll soon update kernels of other umls on other two hosts.
12:30<jdike>that's fine
12:36|-|jjkola [] has joined #uml
12:40<dgraves>jjkola: yo!
12:40<dgraves>jdike: i don't know, actually. i'm not a big latex guy.
12:40<jdike>I figured it out
12:41<dgraves>can you?
12:42<jdike>yeah, latex2html includes some extensions which do the right thing
12:42<jdike>\htmladdnormallinkfoot makes a footnote in the dvi and a normal html link in html
12:46<fo0bar>jdike: I've been having some various problems with sleep() and/or clock drift on 2.6.18 (yeah yeah). I think I've found a hack that work for the moment, that involves CONFIG_UML_REAL_TIME_CLOCK=y and short-circuiting do_settimeofday in arch/um/kernel/time.c to just return 0 instead of doing anything
12:46<fo0bar>thus hopefully locking the time to the host
12:47<jdike>it normally is locked to the host
12:47<jdike>how did fiddling settimeofday change anything?
12:47<jdike>IOW, who is changing the clock?
12:48<jdike>UML_REAL_TIME_CLOCK should be on normally
12:48<fo0bar>I'm able to do date -s on a normal UML guest and local time can be set to a different time than host
12:49<fo0bar>should that not be the case?
12:49<jdike>by "locked", I meant in the absence of someone actually changing the time inside UML
12:50<jdike>so, what exactly is the problem?
12:51<fo0bar>problem is when I set the guest time to something in the past (which shouldn't be possible as I just found), say 5 seconds in the past, then do a "time sleep 1", wall is 6s
12:52<fo0bar>similarly, time set in the future (relative to the host) makes sleep not actually sleep
12:52<jdike>I would try a more recent UML
12:52<jdike>there have been fixes in that area
12:52<fo0bar>why did I guess you would say that :)
12:53<jdike>there a couple of bugs that resulted in sleep not sleeping for the correct time
12:54|-|hfb [] has joined #uml
12:54<fo0bar>I've been hoping to stay with 2.6.18 and backporting fixes as needed because I don't want to keep upgrading 50 guests or so every time a new kernel comes out. by using debian's 2.6.18 as a source base, I can just update when new security patches are released
12:54<fo0bar>but yeah, I plan on checking out .22 and figuring out what's different
12:55<jdike>you might look at the UML patches that go into -stable for a good approximation of what should be backported
12:55<fo0bar>-stable as in your source tree?
12:56<jdike>as in 2.6.22.x
12:56<jdike>if I create and fix a bug in the same release, the fix obviously won't make it to -stable
12:57<jdike>but any serious bug that crosses releases, I try to send to -stable
13:09|-|dgraves [] has quit [Remote host closed the connection]
13:12|-|dgraves [] has joined #uml
13:31|-|tyler [] has quit [Remote host closed the connection]
14:34<kokoko1>wonders its oracle proxy?
14:43|-|nessie [] has quit [Ping timeout: 480 seconds]
15:08<dgraves>kokoko1: nope, the network went poof!
15:10<kokoko1> <-- some nice comments ;)
15:23<jdike>and off it goes
15:27<kokoko1>goes to mainline?
15:31<jdike>paper off to conference
15:32|-|mgross [] has joined #uml
15:32<jdike>Hey mark
15:43<kokoko1>is this okay ..
15:44<kokoko1>Aug 1 13:58:31 shiwi kernel: line_ioctl: tty0: ioctl TIOCLINUX called
15:44<kokoko1>Aug 1 13:58:34 shiwi last message repeated 31 times
15:44<kokoko1>on uml?
15:47<kokoko1>gotta run now, later
15:53<mgross>hay jeff.
15:55<mgross>I'm just lurking on the IRC while I get the hang of UML for some kernel development.
15:55<mgross>Most things are working ok, I'm not yet happy with getting the network up within my uml yet.
15:56<mgross>but I haven't given it a good enough try to bother asking anyone yet.
15:57<mgross>BTW has anyone else had problems booting the "Ubuntu-FeistyFawn-i386-root_fs? (it hangs in early boot, where the Debian 4 fs boots up fine)
15:58<jdike>what version of UML?
15:59<mgross>2.6.22 ? stock lkml.
16:00<mgross>the same kernel boots the deb 4 fs.
16:01[~]jdike downloads
16:02<mgross> is the one I've had issues with.
16:03<jdike>yup, got it
16:04<jdike>.23-rc1-mm1 boots here
16:04<jdike>where does it hang?
16:06<mgross>damn. I feel dumb. it hangs if I include mem=128M, it boots if I leave off the mem= all together. :(
16:06<jdike>well, that's still not good
16:06<mgross>any guess on how much the Ubuntu needs to work?
16:06<jdike>they should all boot with default setting
16:06<jdike>well, I just booted with defaults
16:07<jdike>and the default is 32M
16:07<jdike>it can't hurt to give it more memory
16:07|-|HuK0B [~HuK0B@] has quit [Ping timeout: 480 seconds]
16:07<mgross>hmm. 512M isn't enough. (which is crazy, I think there is somthing up with the mem= )
16:08<jdike>you booted with 32M
16:08<jdike>so "enough" isn't the issue
16:08<mgross>ok, now its not booting without the mem=, after it did just a min ago.
16:08<jdike>something else is going on
16:08<jdike>where is it hanging?
16:09<mgross> * Setting preliminary keymap...
16:10<mgross>this time it hangs at: Couldnt get a file descriptor referring to the console
16:10<jdike>attach gdb to it and see what's going on
16:11<mgross>I just killall linux, now i have some zombi consoles.
16:21<mgross>which linux do I attach too? ps -aux | grep linux shows 8 possible PIDS (4 pids with the same uml invocation command line, and 4 that are just "[linux]"
16:21<jdike>the lowest pid
16:21|-|nessie [] has joined #uml
16:23<mgross>bt shows i'm in start_kernel/rest_init/cpu_idle/init_idle_skas/default_idle/idle_sleep/nanosleep/__kernel_vsyscall
16:24<jdike>where did the boot get to?
16:24<mgross> * Setting preliminary keymap...
16:24<jdike>is that consistent?
16:24<mgross>and one xterm virtual console.
16:25<mgross>I'm not the right guy to say its consistent. I'm sort of fumbling around at this point.
16:25<jdike>try uml_mconsole <umid> sysrq t
16:28<mgross>umid == pid?
16:29<jdike>look for
16:29<jdike>mconsole (version 2) initialized on /home/jdike/.uml/7exryG/mconsole
16:29<jdike>and in this case, the umid is 7exryG
16:31<mgross>in my case the umlid is "NOKimy" and the command "uml_mconsol NOKimy sysrq t, resulted in the Ubuntu umle to boot up.
16:32<jdike>and no output?
16:32<mgross>Memory: 30264k available ??? 30 MB is the default UML ram allocation?
16:32<jdike>32M actually
16:33<mgross>I do have some output.
16:33<mgross>line_ioctl: tty1: ioctl KDGKBTYPE called
16:33<mgross>line_ioctl: tty0: ioctl KDGKBTYPE called
16:33<mgross>line_ioctl: tty0: ioctl KDGKBTYPE called
16:33<mgross>line_ioctl: tty1: ioctl KDGKBTYPE called
16:33<mgross>line_ioctl: tty0: ioctl KDGKBTYPE called
16:33<mgross>line_ioctl: tty0: ioctl KDGKBTYPE called
16:33<mgross>Couldnt get a file descriptor referring to the console
16:33<mgross>a few time and then it finished booting up.
16:34<mgross>now that its booted up I get "ERR Sysrq not compiled in
16:34<mgross>" duh....
16:34<mgross>no output.
16:37<jdike>OK turn on CONFIG_MAGIC_SYSRQ and try again
16:37<mgross>I'm trying from the menuconfig and I'm not seeing the sys-key entry under debugging.
16:38<jdike>it's under UML-specific options
16:43<mgross>should I be concerned about this warning? :WARNING: vmlinux(.bss+0x81fa604): Section mismatch: reference to .init.text:do_pre_smp_initcalls (between 'stdout@@GLIBC_2.0' and 'completed.5758')
16:45<mgross>now I have 5 virtual consoles and they are all some what hung. with the boot up console showing : line_ioctl: tty4: ioctl KDGKBTYPE called
16:45<mgross>line_ioctl: tty0: ioctl KDGKBTYPE called
16:45<mgross>line_ioctl: tty0: ioctl KDGKBTYPE called
16:45<mgross>line_ioctl: tty4: ioctl KDGKBTYPE called
16:45<mgross>line_ioctl: tty0: ioctl KDGKBTYPE called
16:45<mgross>line_ioctl: tty0: ioctl KDGKBTYPE called
16:45<mgross>Couldnt get a file descriptor referring to the console
16:45<mgross>attempting syskey.
16:46<mgross>that kicked things again and its booted.
16:47<jdike>and what kind of output did the sysrq give you?
16:47<mgross>OK SysRq : Show State
16:47<mgross> free sibling
16:47<mgross> task PC stack pid father child younger older
16:47<mgross>init S FFFFE410 0 1 0 (NOTLB)
16:47<mgross>00000001 00000000 08232a60 08058289 08231b1c 08237b3c 08232000 08232000
16:47<mgross> 08232a80 08055bfe 08231620 08237640 08232000 087e1e10 08231620 08237640
16:47<mgross> 08232ac4 08187974 08231620 08237640 08231620 080649c7 00000001 08231620 Call Trace:
16:47<mgross>08232a4c: [<08058289>] switch_to_skas+0x3d/0x6e
16:47<mgross>08232a64: [<08055bfe>] _switch_to+0x4f/0x8a
16:47<mgross>08232a84: [<08187974>] schedule+0x3ea/0x456
16:47<mgross>08232ac8: [<08188030>] schedule_timeout+0x17/0x9c
16:47<mgross>08232af8: [<080af74c>] do_select+0x365/0x3c1
16:47<mgross>08232d98: [<080afa08>] core_sys_select+0x260/0x27d
16:47<mgross>08232ee0: [<080afaea>] sys_select+0xc5/0x16a
16:47<jdike>don't paste it all here
16:47<mgross>08232f34: [<0805868e>] handle_syscall+0x86/0xa0
16:47<mgross>08232f80: [<08066c08>] handle_trap+0xd8/0xe1
16:47<mgross>08232f9c: [<08066ffb>] userspace+0x138/0x180
16:47<mgross>08232fe8: [<0805832c>] new_thread_handler+0x72/0x7e
16:47<mgross>08232ffc: [<a55a5a5a>] 0xa55a5a5a
16:47<jdike>dump it in or something
16:51<jdike>08a12be8: [<08187a4b>] wait_for_completion+0x6b/0x9c
16:51<jdike>08a12c10: [<080615f9>] xterm_fd+0xa7/0xce
16:51<jdike>08a12c3c: [<0806144b>] xterm_open+0x18c/0x20b
16:51<jdike>08a12d94: [<08059446>] open_one_chan+0x40/0x55
16:51<jdike>08a12db4: [<080594d5>] enable_chan+0x13/0x52
16:51<jdike>08a12dc8: [<0805a750>] line_open+0x33/0x8f
16:51<jdike>08a12ddc: [<0805900c>] con_open+0x10/0x12
16:51<jdike>08a12dec: [<0812deff>] tty_open+0x123/0x26c
16:51<jdike>an xterm was failing to open properly
16:51<jdike>add "con0=fd:0,fd:1 con=pts" to your command line?
16:53<mgross>now it boots but I don't get any virtual console windows.
16:53<mgross>host problem with my xterms?
16:53<jdike>I guess
16:53<jdike>might be a race inside UML
16:54<mgross>My host is i386 Ubuntu 7.04, if that helps at all
17:00|-|dang [~dang@] has quit [Quit: Leaving.]
17:00|-|dang [~dang@] has joined #uml
17:01|-|dang [~dang@] has left #uml []
17:13<mgross>hay, thanks for the help ! I'll be back later if I get into trouble with getting network access from within the UML.
17:19<mgross>one more question , about sysrq m
17:20<mgross>it seems that uml_mconsol <umid> sysrq m doesn't produce output after the uml is booted.
17:20<mgross>Where does it dump its console messages?
17:20<mgross>the data is in the uml dmesg however.
17:54|-|ram [] has quit [Ping timeout: 480 seconds]
18:16|-|hfb [] has quit [Quit: Leaving]
18:23|-|ram [] has joined #uml
18:31|-|kokoko1 [~Slacker@] has quit [Ping timeout: 480 seconds]
18:31|-|mgross [] has quit [Quit: Leaving]
19:17|-|horst [] has quit [Remote host closed the connection]
20:46|-|jjkola [] has quit [Quit: *pop*]
21:13|-|ram [] has quit [Ping timeout: 480 seconds]
21:44|-|ram [] has joined #uml
21:53|-|ram_ [] has joined #uml
21:53|-|ram [] has quit [Read error: Connection reset by peer]
22:35|-|jdike [] has quit [Quit: Leaving]
22:59|-|VS_ChanLog [] has left #uml [Rotating Logs]
22:59|-|VS_ChanLog [] has joined #uml
23:23|-|ram [] has joined #uml
23:23|-|ram_ [] has quit [Read error: Connection reset by peer]
---Logclosed Thu Aug 02 00:00:18 2007