/srv/irclogs.ubuntu.com/2009/07/10/#ubuntu-kernel.txt

dtchenKano: ar9170usb isn't enabled in the Ubuntu builds AFAIR; i've been using compat-wireless with the firmware from linux-firmware.git.00:52
Kanoyou are lying00:52
dtchenis it enabled now for karmic?00:53
Kanohttp://packages.ubuntu.com/search?searchon=contents&keywords=ar9170usb.ko&mode=exactfilename&suite=karmic&arch=any00:53
Kanothey are of course in the image00:54
dtchenah, good00:54
dtchenapw: would you sync the ar9170usb config (=m) in the c-o-d builds, too?00:55
Kanowhere are the 2.6.30 source packages00:55
dtchencanonically? they're gone; karmic is using 2.6.31-git. you'll have to use launchpadlibrarian for older source packages (e.g., https://edge.launchpad.net/ubuntu/+source/linux/2.6.30-10.12).00:57
Kanoyou dont used a orig file?00:58
dtchenit doesn't appear so00:59
Kanono errors about dm with 2.6.31 in launchpad?01:03
Kanothat's definitely broken01:03
dtcheni can't speak for karmic's source package at the moment; i'm using a c-o-d build (2009-07-09)01:05
Kanoc-o-d?01:07
dtchen"crack of the day" builds: http://kernel.ubuntu.com/~kernel-ppa/mainline/daily/01:11
Kanoi see01:12
Kanowenn i can not use those, i do not use ubuntu as base. i always have to apply some patches before building01:12
dtchenare you patching pristine upstream? those are upstream kernel.org sources.01:13
Kanono some config settings01:15
Kanoi enable the wireless legacy code and some other things on the fly, also i correct known bugs01:16
Kanosomehow u likes to have some bugs, prefer to fix em01:16
dtchen:-)01:17
Kanolike my kernel has working dm, comedi s626 only loading for one special device not generic, and some other changes01:18
Kanohttp://kanotix.com/files/excalibur/linux-2.6.31-generic/source/01:19
Kanoi even have got aufs201:19
Kanoalso i use lzma compression01:20
Kanofor vmlinuz01:20
sn9which lzma patch did you choose? there are several01:26
Kanono patch01:27
Kanothats one config setting01:27
KanoCONFIG_KERNEL_LZMA=y01:28
Kanoi compress kernel + initrd with lzma, u uses lzma compression for initrd only in live mode it seems01:30
Kanobut they still did not compress the kernel image with lzma01:30
apwsmb it is possible that it can be allowed to be interrupted08:43
apwas the semantic is likely 'when complete all cpus must have flushed the tlb'08:43
smbHm, ok. Cause in that one case it looked like08:43
smbfrom the printk's I saw the printk before calling the function (after changing the attrs) and the BUG trace is from some acpi code08:44
smbbut I have two more printk's in global_flush_tlb08:44
apwthere is no guarentee they have not been 'printed' and not yet flushed out08:45
apwwhich acpi code is blammoing08:45
smbwouldn't it with the BUG statement? Something generic called from thread helper: autoremove.wake.function->...>acpi.os.execute.deferred08:46
apwdepends on the output device, all printk does is queue it for printing in the dmesg buffer and then wake up the drivers08:48
apwsome drivers are more immediate than others08:48
apwserial for instance definatly is not08:49
apwsmb what is the specific panic, do you have a picture?08:50
smbHm, currently using vga with smal font08:50
smbapw, yes08:50
apwthere must be more information in there, can you get the picture up somewhere08:50
smbnull pointer dereference08:50
apwhrm08:50
apwif the remapping was gone correctly then it would not make any difference if the flush has occured or not08:52
smbhttp://www.krikkit.de/aa1debug/07100001.JPG08:54
smbhttp://www.krikkit.de/aa1debug/07100003.JPG is a bit better08:54
smbhttps://pastebin.canonical.com/19558/08:59
apwthat is bugging twice08:59
smbThat would be the code snippled called08:59
apwthe stack is not from the original bug by the looks08:59
apwand the first one is between your two debug statements08:59
smbapw, Oh it is only one. Just two pictures of it09:00
apwneed to find our where that first EIP is on line 409:00
smbBut the rest yeah09:00
apwsmb, no in the second picture there are two panics09:00
apwone on lines 4-609:00
apwand one on 7-09:01
apwthe stack is as likely to be from the second one09:01
smbOh, right. The Null pointer and then Oops: 000309:01
apwneed to translate c0105550 on your kernel09:01
smbJust a second. Booting the build machine09:02
smbHm, c0105550 -> device_not_available09:04
smbapw, Oh actually I just noticed the between printing the eip of the first oops and the BUG, the flush_map(...) is from within global_flush_tlb09:12
smbOne of the debug statements I added09:12
apwyeah saw that09:12
smbHrm, missed it on the camera and on the screen :/09:13
apware interrupts turned off during the conversion to r/o?09:14
apwdo you know if the other cpu threads are awake when that code runs?09:15
smba) I don't think so, just spinlocks while taking lists out afaik. b) not sure but checking09:15
apwsmb, ok the first panic is very odd as its panicing on the stack push09:16
smbb) it is after bringing up the hyperthreads09:17
smbapw, dumb question, how do I see this?09:17
apwso its possible the other thread is losing its stack or similar09:17
apwENTRY(device_not_available)09:18
apw        RING0_INT_FRAME09:18
apw        pushl $-1                       # mark this as an int09:18
apw        CFI_ADJUST_CFA_OFFSET 409:18
apw        pushl $do_device_not_available09:18
apw        CFI_ADJUST_CFA_OFFSET 409:18
apw        jmp error_code09:18
apw        CFI_ENDPROC09:18
apwEND(device_not_available)09:18
apwit the eip is exactly device_not_available ... then it is pushing a frame09:18
smbOk, yes it is. As the next eip is exactly the entry of call_function_interrupt09:19
apwsmb, i don't understand09:19
apwphone?09:19
smbok09:20
* apw watches his fan turn on and off correctly on his thinkpad .. amazing11:08
sahil_hey all, got a bit of an issue, when i boot with acpi=off everything boots but of course acpi does not work, when i boot with nolapic acpi works but wireless does not, any ideas?13:12
apwwhy are you booting with either of those?13:19
=== fabbione is now known as fabbione-vac
primes2hogasawara_: I have again the laptop I told you some days ago. Tell me if you need something.14:38
primes2hogasawara_: Ehm, Hello, first of all... ;)14:47
maco2.6.31-2: does it include the fix for tun interfaces?14:53
rgzIs there a generic problem with the 2.6.28-13 kernel build on i386?15:06
smbrgz, It runs for me15:10
smbWhat problem do you have?15:10
rgzsmb:  It first gives off some error about my resolution bindings, after grub's loaded; says that 307 doesn't exist, I change the splash resolution, it boots up gdm but all I/O devices don't function.15:13
rgzI'll get a log as soon as I can, just wondered if anyone else experienced something similar.15:14
smbrgz, It does not ring a bell immediately. Hm, 307 might be a resolution but why this affects IO devices. HAve you once tried removing splash?15:16
rgzsmb:  Yeah, I looked through the logs, might be something to do with nvidia drivers, but I'm not entirely sure.15:21
rgzIt's a pain; but I'll come back when I have the logs in check.15:23
smbrgz, Just to enlighten me. ;-) Did using splash or not make a difference. Ok15:23
nanomadCan anyone have a look at #21367 it can be fixed in (at most) 5 minutes15:26
rgzsmb:  No, no difference.15:27
smbrgz, ok thanks15:27
=== Nicke_ is now known as Nicke
jdstrandthere is a newer version of firmware for ipw2200 than what's in linux-firmware (bug #357968). license in the tarball is the same. would I be stepping on anyone's toes if I uploaded an updated linux-firmware to karmic?16:08
ubot3Malone bug 357968 in linux-firmware "Please update ipw2200 firmware to version 3.1" [Undecided,Triaged] https://launchpad.net/bugs/35796816:08
smbjdstrand, I would not believe so. I believe the names were incremental (so old ones stay there), right?16:14
jdstrandsmb: I don't understand your question. it is a native package, I would adjust the version appropriately16:14
smbjdstrand, It is rather about the version of the firmware file. I just believe those get adapted, sod the old firmware files stay16:15
jdstrandsmb: linux-firmware simply has the untarred files in firmware/ipw2200. I figured I'd put the new ones there instead16:17
jdstrandis that not the proper way?16:17
jdstrandall the filenames are the same16:17
jdstrandthe LICENSE is the same16:17
smbOh I see I confused that with other wireless firmaware which has version numbers in the filename and the driver loads just the highest supported16:19
smbSo here it is just a replacement as it looks16:19
* jdstrand nods16:20
smbThere has been a request to upgrade the iwl4965 just today. I wonder whether it will make sense to get that done with one update16:21
smboh.. forgott, jdstrand ^16:22
jdstrandsmb: I don't know anything about iwl4965. I can say preliminary testing shows the ipw2200 driver works fine on 2.6.28 (haven't tested on karmic)16:23
smbapw, Would you feel being stepped on something when the firmware package gets done?16:23
jdstrandI'm happy to do an ipw2200 only upload16:24
jdstrandotherwise, the details are in the bug for ipw220016:24
smbjdstrand, Ok, for simplicity. Lets do one by one16:24
jdstrandalright, consider it done :)16:24
apwsmb, jdstrand the only thing about firmware is legal issues, i would just touch based with pgraner to make sure he is happy16:29
jdstrandapw: the license in the new tarball is the same as in firmware/ipw2200 of linux-firmware16:29
apwthen that sounds fine to me, i would just let pgraner know you are updating it, as there is a review ongoing16:30
pgranerjdstrand: then it good16:30
jdstrandthanks guys-- will test on karmic and upload16:30
pgranerapw, jdstrand: I have it noted. Thanks for the heads up16:30
nanomadjdstrands, Can anyone have a look at #21367 also?16:30
jdstrandnanomad: not really my jurisdiction, sorry16:31
nanomadjdstrand, np, thanks anyway16:31
jdstrandnanomad: and that isn't linux-firmware, bug linux anyway16:32
nanomadyes, i noticed that :)16:32
apwbug #2136716:32
ubot3Malone bug 21367 in linux "Wifi-enabled led is not lit on ipw2200 cards" [Low,Confirmed] https://launchpad.net/bugs/2136716:32
smbThe question is really why upstream did not change the default. It might be work on many but just not on all. Pragmatically I got a ipw2200 and my led works without...16:34
apwnanomad, whats the question with it?16:35
nanomadapw, it got disabled in karmic16:35
apwit was fixed in Jaunty if I understand things16:35
nanomadsince it is not fixed upstream16:35
nanomadit got lost from jaunty -> karmic16:35
nanomadif i manually modprobe ipw2200 with led=1 it works16:36
smbHm, ok. So we had it with the other default in Jaunty16:37
nanomadsmb, looks like that16:37
apwsmb can you accept the nom for jaunty for me on that bug16:37
apwand i'll sort out the states16:37
* smb wait for lp16:38
smbok16:38
nanomadactually, it would be nice to report it upstream. Or we will have the same bug at the next kernel bump16:38
apwsmb, ok thats been applied to jaunty since the karmic fork and all the commits there are on my list to review16:38
smbapw, We should check to get it upstream this time,...16:39
smbnanomad, yep16:39
smbEither we will know why not or wo't loose it agian16:39
apwsmb yep.  will do that as its been on for 3 months16:39
smbapw, ok, quickly scanned the ipw2200 bugs and it did not look like there was one of "suddenly broke"16:43
apwsmb, might be worth rejecting the nom for karmic on that bug as it means nothing16:43
apwsmb, thanks for that16:43
smbapw, Yeah, will reject that. You are not on it nw?16:44
apwi've assigned it me and added it to my list to close when karmic has it16:46
smbapw, ok, nomination rejected16:46
=== bjf_afk is now known as bjf
apwsmb, thanks16:47
nanomadapw, smb, thanks guys.16:49
=== maco_ is now known as maco
jdstrandapw: I uploaded linux-firmware 1.14 to karmic. I tried to git clone http://kernel.ubuntu.com/git-repos/ubuntu/linux-firmware-karmic.git/ but got:17:05
jdstrandfatal: http://kernel.ubuntu.com/git-repos/ubuntu/linux-firmware-karmic.git//info/refs not found: did you run git update-server-info on the server?17:05
apwits meant to get run on every commit magically17:05
jdstrandapw: I uploaded the new source package to /home/jamie/uploads/karmic on chinstrap17:05
jdstrandapw: I looked around for docs on how to update linux-firmware, but nothing popped up17:06
jdstrandapw: if you point me at it, I'll do it, otherwise you can get the source package on chinstrap or wait for LP17:06
apwack17:07
wiphi people, i've been using jaunty for months, but the RT-kernel is still not updated. it's full of bug. any release date? is someone working on it?17:09
wipexample, worst bug for me: https://bugs.launchpad.net/ubuntu/+source/linux-rt/+bug/37402617:12
ubot3Malone bug 374026 in linux-rt "Kernel threads using 100% on one of the two CPUs" [Undecided,Confirmed] 17:12
smbapw, Is there actually a git for linux-firmware...?17:12
apwsmb dunno, wanna try and find out and sort that for me17:13
* apw is snowed17:13
wipstill undecided...17:13
smbwip, The rt kernel is community maintained. Don't know the status17:13
smbapw, can do17:13
wipsmb: thanks for the information, it17:15
wipsmb: but it means that i should recompile my kernel... ouch...17:15
=== ogasawara_ is now known as ogasawara
Q-FUNKhello20:09
Q-FUNKcould anyone help me with LP bug #396286 ?20:10
ubot3Malone bug 396286 in linux "kernel 2.6.31-generic oops after loading initramfs" [Undecided,New] https://launchpad.net/bugs/39628620:10
Q-FUNKapw: I beleive that you were the one who replied to LP bug #241307 earlier.  I think that we managed to narrow down the source of the problem.  is there anythign else I can add?20:12
ubot3Malone bug 241307 in linux "kernel oops during bootup in LTSP" [High,In progress] https://launchpad.net/bugs/24130720:12
apwQ-FUNK, looks like you have some significant update there20:13
apwi need to read what you've said and comment, busy right now, will look next week20:14
Q-FUNKapw: at least, it looks like wwx and leio/mraudsepp (freenode) have done a good job of narrowing down the cause.20:14
Q-FUNKapw: wwx pointed out that with early releases of kernel 2.6.23, it was still possible to revert the change with an easy patch, but not anymore.20:15
Q-FUNKalright.  I'll wait for your followup, then. :)20:15
apwyeah probabally futzed by the merge of 32 and 64 bit x8620:16
apwon my list for next week20:16
Q-FUNKis that what the code that replaced the old ia32 init code in ASM does?  attempt to combine init code for x32 and x64 into one?20:17
apwno i mean that is what likely futses reverting the patch20:20
apwthe previous bug you mention sounds liek a very early panic, normally we take piccies of those with a digitcal camera and attach those to the bug20:20
Q-FUNKI'd gladly take a pic (as I previously did for the 2.6.24 kernel bug), but not much core dump fits on the default 24-line VGA screen.20:21
Q-FUNKyes, it indeed feels like an early kernel panic.  I'm just not sure what causes it20:21
apwif you add vga=ask to the kernel command line it lets you pick a smaller font20:23
Q-FUNKyup.  but the max still is 80x60.  we cannot quite fit the whole kernel dump20:33
Q-FUNKanyhow.  attached.  hope it helps.20:34
vitovtHi everybody! I need more recent kernel like 2.6.29. I installed it from http://kernel.ubuntu.com/~kernel-ppa/. But there is no restricted extras. Where I could get config of existing kernel (like 2.6.28) including RESTRICTED EXTRAS. How could I buld new kernel very similar to default ?20:41
macovitovt, um, restricted extras packags just have like...flash and java and mp320:56
vitovtow. sorry. i mean restricted-modules20:58
vitovtit's too late in my country and I sleeppy now :)20:59
vitovt221:04
vitovt321:04
vitovt421:04
vitovt521:04
vitovt621:04
vitovt721:04
vitovt21:04
vitovtI have a big problem with 2.6.28 kernel in mainline. 21:04
vitovtWhen i'm out of RAM - swap usage allmost impossible.21:04
vitovtThere are a lot of errors in syslog (Error allocating memory).21:04
vitovtI used kernel 2.6.29  from (http://kernel.ubuntu.com/~kernel-ppa/). It works great. But miss some important modules for me. 21:04
vitovtAlso I need to make some modification (framebuffers support)21:04
vitovtHow I can build restricted extras for 2.6.29 and make configuration allmost similar to 2.6.28-generic ?21:04
vitovtSorry for possible flood21:04
dtchenvitovt: clone the git tree; see kernel.ubuntu.com/git/21:45
dtchenvitovt: see the knowledgebase on the kernel wiki for pointers to build it against the appropriate kernel headers21:46

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!