/srv/irclogs.ubuntu.com/2008/01/28/#ubuntu-kernel.txt

joejaxxhello all i redid a patch for bug 162090 against the latest hardy kernel revision i was wondering if someone could review it :D02:37
ubotuLaunchpad bug 162090 in linux-source-2.6.22 "appletouch does not recognize trackpad in macbook3,1" [Undecided,Confirmed] https://launchpad.net/bugs/16209002:37
joejaxxit is a git diff :)02:37
=== ln-_ is now known as ln-
tonyyarussoCould someone please tell us at least a temporary workaround to https://bugs.edge.launchpad.net/ubuntu/hardy/+source/linux/+bug/129910 ?  The things people have found so far only work some of the time, so are clearly missing something.03:44
ubotuLaunchpad bug 129910 in xserver-xorg-video-ati "Blank ttys when using vesafb (vga=xxx)" [Unknown,Fix released] 03:45
krautmoin08:12
=== doko_ is now known as doko
Kanohi, i added 2 patches to09:07
Kanohttps://bugs.launchpad.net/ubuntu/+source/linux/+bug/18571909:07
ubotuLaunchpad bug 185719 in linux "remove duplicate pci ids bcm43xx/ssb prism54/p54pci" [Undecided,New] 09:07
Kanothey are tested and working09:07
Kanodont forget the firmware (b43-fwcutter)09:29
Kanolaga: did you try aufs?10:42
Kanoanybody awake11:51
gicmoI am11:51
gicmobut I am not sure of I am of any help ;-)11:51
Kanoi just want those patches integrated11:52
Kanoand best: addone of current (28.1.) aufs to lum11:52
Kanonew code since today11:52
gicmoI have nothing to to with the kernel team ;-)11:53
=== asac is now known as asac_resurrected
=== asac_resurrected is now known as asac
=== lamont` is now known as lamont
=== \sh_away is now known as \sh
rtgabogani: I'm not registered so can't reply. Can you put a patch on zinc somewhere?14:06
aboganirtg: Sorry i don't know that you aren't registered.14:12
aboganirtg: Sure! Please check it at my home on Zinc (/home/abogani/1-rt-compat-symbols.diff).14:12
aboganirtg: Only one note: i don't have idea if CONFIG_XEN is still necessary.14:12
rtgabogani: no problem. I chose not to register since the conversations regarding kernel stuff should be in the open.14:12
rtgabogani: your patch is a good example of why lrm should be under git revision control. However, the size of the binary blobs makes it somewhat unmanageable.14:14
rtgabogani: FYI - Xen has not been enabled yet.14:14
aboganiAgreed14:16
zulrtg: patch should be coming today14:42
rtgzul: good timing since I totally fubared the -5.8 upload.14:43
zuloh?14:43
rtgI rebased against 2.6.24, but somehow missed the ABI check failiure.14:43
zulheh oops :)14:43
rtganyway, -6.10 will be the next upload.14:43
rtgzul: correctiom, it was the -5.9 upload that fialed.14:44
zulheh i have a patch for lum as well so Ill be doing that one first (ie sending email to kernel team list yadda yadda)14:45
rtgzul: be sure to check the xen build against -6.10. I'm done rebasing for awhile.14:45
sorenrtg: So, I might still be able to get something shoved into the kernel before alpha 4?14:47
rtgsoren: dunno yet if the -6.10 upload will get cooked in time.14:48
rtgsoren: send me you pull request anyway.14:48
sorenrtg: Will do.14:49
lagartg: any news wrt unionfs?15:39
rtglaga: refresh my memory.15:44
lagartg: sure. there was supposed to be a new l-u-m upload with a unionfs patch which fixes bug #10304415:45
ubotuLaunchpad bug 103044 in linux-ubuntu-modules-2.6.22 "kernel oops during boot on a nfs root diskless system, SRU TEST CASE" [High,Confirmed] https://launchpad.net/bugs/10304415:45
rtglaga: likely this week. kees is still testing...15:46
lagartg: that's good news. i'll ask him if i can have the patch15:47
rtglaga: we were both traveling last week, so there has been disruption in our workflow.15:47
lagayes, i figured as much. 15:47
rtglaga: you should be able to get what you need from the git repo.15:47
lagaoh. i haven't seen any related commits. guess i should look again.15:47
rtglaga: i just read the bug report. I don't think the problem you are seeing is solved.15:49
lagaoh. :/15:49
sorenrtg: pull request sent.15:49
lagawhich would explain why i can't find the commit.15:49
lagabecause i already have the latest l-u-m from gutsy-proposed.15:50
lagartg: any chance it will get solved?15:51
rtglaga: do you know of an upstream patch for it?15:51
lagartg: no. i didn't ask upstream because ubuntu still has 1.4 which is quite old15:52
rtglaga: if you can find and test one, I'll consider it. An oops satisfies SRU policy.15:52
rtglaga: otherwise your solution may be Hardy.15:53
lagartg: i do not believe that hardy solves my problem.15:53
lagathis setup on hardy still breaks, i just never got around to look at the OOPS15:53
rtglaga: we have a better chance of getting upstream to look at it.15:54
rtgin Hardy that is15:54
lagato be honest, i just need a working stacking file system in hardy. gutsy doesn't matter.15:54
lagartg: oh. does hardy have newer unionfs?15:54
rtglaga: ok, I'll have -6.10 uploaded soon which is based on -6.10. unionfs looks like it is still 1.415:58
lagartg: yes, that's why i was asking :)15:58
rtglaga: perhaps you could make a patch for current upstream unionfs and test it against Hardy?15:59
lagartg: i need a stacking file system for Mythbuntu - people can have diskless fat clients that way. if unionfs can't be fixed, aufs ("another unionfs") might be an alternative. it requires a very small kernel patch (just a few symbol exports) to get NFS working, though.15:59
lagartg: you mean i should build current unionfs and test it in hardy? what about the apparmor changes needed for unionfs in gutsy?16:00
rtglaga: since AA is not in the kernel, I imagine upstream unionfs will have those fixes. dunno for sure.16:01
rtgs/not/now/16:01
lagaoh.16:01
lagai didn't know that apparmor made it in the kernel. that's good news.16:02
lagartg: if my testing was successful, would you add current unionfs to hardy or just cherry-pick patcheS?16:02
rtglaga: I would probably add current unionfs. It will get some scrutiny however, given the problems we've had in the past.16:04
lagartg: yes.16:04
lagathought as much :)16:05
lagartg: would adding aufs be an option? that'd mean there was another stacking filesystem which also works with NFS while keeping unionfs in a known, stable-enough state16:06
rtglaga: yep.16:10
rtglaga: it should go into ubuntu-hardy-lum/ubuntu/fs16:11
lagartg: i've just remembered that upstream aufs has suggested some patches which would need to be applied to the kernel. ogra told me you (as in "the kernel guys") wouldn't do it which would make adding aufs pointless16:12
rtglaga: it depends on what the patches are. I'd have to see them first.16:12
lagalet's see...16:13
lagartg: here's patch #1:16:18
lagahttp://aufs.cvs.sourceforge.net/aufs/aufs/patch/sec_perm-2.6.24.patch?revision=1.1&view=markup16:18
laga#2: http://aufs.cvs.sourceforge.net/aufs/aufs/patch/splice-2.6.23.patch?revision=1.1&view=markup16:19
laga#3 http://aufs.cvs.sourceforge.net/aufs/aufs/patch/put_filp.patch?revision=1.1&view=markup16:19
laga#4 http://aufs.cvs.sourceforge.net/aufs/aufs/patch/deny_write_access.patch?revision=1.1&view=markup16:20
lagathat's all he has mentioned.16:20
rtglaga: I'll look at 'em in a bit. I have some other stuff I gotta work on.16:21
lagartg: sure. me too, actually, so i'll stop procrastinating here :)16:21
=== _MMA1 is now known as MetalMusicAddict
=== MetalMusicAddict is now known as _MMA_
Kanohi rtg 17:35
lagaKano: no, i haven't tried your aufs yet. it's possible that it goes to l-u-m though17:36
Kanowell i did not see a big difference in todays snapshot17:36
Kanortg: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/18571917:37
ubotuLaunchpad bug 185719 in linux "remove duplicate pci ids bcm43xx/ssb prism54/p54pci" [Undecided,New] 17:37
rtgKano: its on my list17:37
Kanothen apply it17:37
rtgKano: patience17:38
Kanothats nothing i have got17:39
Kanolaga: same position, updated cvs17:40
Kanobtw. you should be able to use my precompiled kernels too17:41
Kanohttp://kanotix.com/files/kernel/kernel-update-pack-generic.tar.gz17:42
Kanoin source dir all patches17:42
Kanoif you want to compile on your own use em17:42
Kanoor just use all from kernel + extra (maybe without avm as you dont have the dependency for that)17:42
lagaKano: thanks, probably won't have time before the weekend to do anything17:43
Kanoi can you give already made images using live-helper too17:44
Kanoif needed 2 with same settings, one unionfs next aufs17:44
lagaKano: i don't need live disks, i just need a working stacking file system in the ubuntu repos :)17:45
Kanowell the aufs source does not compile directly17:45
Kanoas it needs at least 2 extra files17:45
lagaKano: the necessary patches are in ubuntu now, the aufs-source maintainer has updated it. he even used one of your patches :)17:46
Kanodid not see the export17:46
Kanowhere is it?17:46
Kanolast weekend all applied17:46
lagaoh17:46
lagai was talking about the aufs-source package17:47
laganothing in git yet17:47
Kanowell i tested a kernel with all patches proposed, did not see any difference compared to applying only one17:47
lagaKano: do you realize you need to configure the aufs module accordingly?17:48
Kanoi just copy 2 extra files to my headers and compile aufs 17:49
lagathere are some defines which toggle those settings17:49
Kanowell i just use autodetection that way17:50
lagait's documented somewhere in the top-level directory of the aufs source tree. don't have it here on my laptop17:50
lagathere's no auto-detection IIRC, just defaults. i could be wrong, though17:50
Kanohttp://kanotix.com/files/kernel/kernel-update-pack-generic/source/17:50
Kanohttp://kanotix.com/files/kernel/kernel-update-pack-generic/source/sec_perm-2.6.24.patch17:51
Kanothats the absolute minimum17:51
Kanoother way would be17:51
Kanodont use that patch and compile aufs STATICALLY as kernel patch17:51
Kanosame would work for unionfs too btw17:56
=== fabbione is now known as thegodfather
=== asac_ is now known as asac
Kanobtw. when i use unionfs, i can boot with live-helper but network-manager does not run18:41
Kanoit works with aufs however18:41
Kanomaybe thats why u does not use networkmanager by default and uses stupid dhcp entries in /etc/network/interfaces...18:44
=== smb is now known as smb_away
=== _MMA1 is now known as _MMA_
=== _MMA1 is now known as _MMA_
=== \sh is now known as \sh_away
=== smb_away is now known as smb

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