=== _LibertyZero is now known as LibertyZero [07:44] apw: just fyi, I'll cover us at the release meeting tomorrow. gcc should just about be done building by then so I plan to do one last kernel upload to pull in what we have in master-next for Alpha-1. [07:49] hi [07:50] after upgrading to ubuntu 11.04 (for 10.10) i have troubles with my SD card reader integrated in acer extensa 5635 [07:50] simply put SD card in usb device (some realtek reader ) is recognized but then after some seconds is disconnected === fairuz_ is now known as fairuz [07:50] very strange [07:51] works fine with 2.6.35 [07:51] any ideas or it is known pb? [08:05] open-nandra, nothing i've heard about no [08:05] pls file a bug against linux === smb` is now known as smb [08:24] morning [08:24] morning smb [08:25] morning [08:25] morning [08:26] morning: ppisati, apw [08:26] jjohansen, late one for you [08:27] heh yeah, Ian had a late nap, so he will be up for at least a couple more hours so I am writing [08:27] code or documentation? [08:27] smb: documentation after a fashion, its a paper [08:29] anyone else running natty with an external monitor, and if so, are the indicators on the external monitor creaping off the right slowly over time? [08:30] apw, Well I run it on my desktop... [08:30] this is over days i am noticing it moving off, the power button me menu and clock have gone now, and a few pixels of the envelope now [08:30] apw: I get indicators being half drawn/covered and indicators disappearing, not seen going off the right [08:31] But the indicators stay I think (if they show up at all, though that could have been because when booting there was an external networking issue) [08:31] yestrerday i only had the date, no time from the time app, now the date is gone as well [08:32] are any of the migration applets in the empty space [08:32] apw: fun, have you tried killing compiz, and or unity to see if it restores them [08:32] Hrm, only got the time there ... and the settings thre brings up ... nothing [08:33] er if it's unity then it's not a real panel and nevermind me :] [08:33] apw: I do have an external monitor, but I switch between it and the laptop's LCD a couple of times a day, and I don't see this happen at all. [08:33] apw: That's an *awesome* bug. That I've not experienced with my external monitor and oneiric. [08:34] soren, yeah i suspect if i unplug and replug it'll get fixed [08:34] apw: If only there was a way to find out... [08:34] the one on the LCD is fine, its just my external [08:34] :p [08:34] Great, no changing time and date settings today... at least not with the ui [08:50] My kubuntu's performance is getting poor when I try to run my laptop on battery. Is there any way to improve it. I tried everything from power management and everywhere but, its not improving. I think its a kernel issue. Can anybody please help? [08:52] c2tarun: that is rather vague, you need to collect more information [08:52] jjohansen: what kind of information? [08:54] c2tarun: well, some firm performance figures showing how things are slower would be good [08:54] a bug with logs, etc [08:55] maybe some runs with powertop, latencytop, .. [08:55] jjohansen: hmm... how can I get log for getting system slow? I can feel it, opening any window and closing it, switching workspaces, openening widgets, everything is getting slower. :( [08:58] c2tarun: right, but your feel of things getting slower isn't something we can debug against [08:59] filing a bug, that collects the logs and system information gives some basic information to start from. [08:59] jjohansen: yeah, you are right, actually I was looking for something that can enhance my system's performance and not fix a bug. [08:59] c2tarun: do you see intel_idle in the output of lsmod? [09:01] ohsix: nope here is pastebin for my lsmod http://paste.kde.org/75217/ === htorque_ is now known as htorque [10:09] g'night [10:11] ppisati, how goes your cve fun [10:12] apw: it goes :) [10:12] apw: i'm piling up the patches [10:12] ppisati, heh, how far through are you? [10:14] apw, If you do not hear us you are buggy [10:14] arrg [10:14] now i hear [10:20] yep [10:47] apw: down 15, i'm applying all of them now [10:48] ppisati, nice ... that'll help out stats [10:48] apw: yep [10:48] apw: i really didn't know of the existence of that page... [10:49] ppisati, heh no worries, that page is really the stable teams input for cves [10:49] and until karmic went dead fsl-imx51 at least was a rebase onto that, now its not we need some other solution [10:50] and you are the solution there, so its a new requirement just since release [10:50] so you not knowing isn't supprising [11:08] kernel is compiling [11:08] * ppisati rushes out to het some food [11:13] hi friend of mine has a 10.04 lts server with ecc ram and gets spammed by edac mc0 errors is there any solution [11:13] ghostcube, Yeah, better RAM [11:13] heh [11:14] But really, had the same and had to find the module that was faulty. [11:15] Those messages tell you that there is a problem. Usually like correctable ecc errors. There is some hints about the location too, but it can be a bit tedious to map them to the physical slots [11:17] seems to be an edac bug [11:17] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/367774 [11:17] Launchpad bug 367774 in linux "EDAC spam in dmesg, edac-utils shows no erros" [Undecided,Expired] [11:17] his ram isnt faulty he testet it [11:30] ghostcube, in that bug all of the users seem to have indicated that quick boot was not clearing the memory out leading to the eec errors [11:30] somone has to write to all of ram before eec can be safely enabled [11:35] yeah he tried this, but its still spamming errors its strange [11:35] one hint is just to blacklist edac modul [11:35] yep stops reporting but they are still there i am sure [11:36] nope, he tried debian 6 hdd and bootet in slow boot now no more errors [11:36] hmm [11:37] and does that kernel even have the edac module for his h/w ? [11:37] they are quite new in some cases [11:37] its a asus pb5v [11:38] p5bv [11:40] http://buttersideup.com/edacwiki/Uninitialized_ECC_bits found this [11:42] https://bugzilla.redhat.com/show_bug.cgi?id=564274 and this [11:42] bugzilla.redhat.com bug 564274 in kernel "fake EDAC errors on i3210" [High,Closed: notabug] [11:42] ok seems to be a bios prob [11:43] ghostcube, thats what it sounds like [12:14] could somebody advise me on what to do about bug 785394? reserving 128MiB of RAM seems like an awful lot, and I wonder why the kdump kernel is taking quite so much memory to boot [12:14] Launchpad bug 785394 in grub2 "Hard-coded crashkernel=... memory reservation in /etc/grub.d/10_linux is insufficient" [Undecided,New] https://launchpad.net/bugs/785394 [12:23] It looks like it's OOMing while unpacking the initramfs ... hmm [12:23] $ zcat /boot/initrd.img-2.6.38-9-generic | wc -c [12:23] 44324864 [12:35] cjwatson, the kernel en-toto is pretty huge, initramfs coming straight out of ram, both compressed and uncompressed [12:36] cjwatson, do we use a standard kernel for the kexec kernel? ie. do we not strip the initramfs at all [12:38] I was hoping one of you lot would know [12:39] i can have a look [12:39] it does have to mount the rootfs, so it seems to me that it would have many of the same requirements [12:39] cjwatson, and the complaint there is that it won't fit into 64M in that bug [12:39] though maybe we could use MODULES=dep for the kdump initramfs, if it's possible for it to be different [12:40] which with a 40M initrd, ... its not going to i suspect [12:40] cjwatson, right that was what i was thinkng, you kexec initrd really doesn't need to be portable [12:41] I don't know if it's possible for it to be separate, though [12:41] cjwatson, me either, shall i take a look ? [12:41] if you could, that'd be great [13:08] lucid -proposed is stuck due to ec2 bug, right? [13:08] i mean linux (2.6.32-32.62) [13:08] https://launchpad.net/ubuntu/lucid/+source/linux/+changelog [13:36] ppisati: yes, was put on hold because of ec2 issue [13:44] herton: i see [14:29] cjwatson, so its is a simple configuration to change which kernel kdump uses, defaulting to the current kernel, whats not so obvious is how we might maintain more than one initrd for a kernel via the normal mechanisms. i don't know if we could consider having on fixed kernel for kdump, as using the likely crashing one to dump itself is a little scarey [15:11] * apw pops out to collect a bag [16:37] what was the script that updated the changelog? [16:37] ppisati: debian/rules insertchanges ? [16:38] herton: i mean, the kteam tools [16:38] herton: i think it's startnewrelease [16:38] I'm going to take a bit of an extended lunch. need to run some errands and check on my truck in the shop. [16:38] may not be gone very long, but wanted to let you know anyway :) [16:38] ppisati: ah sorry, should be maintscripts/maint-startnewrelease [16:39] herton: yes, that one [16:52] herton: but it doesn't update the changelog [16:52] uhm [16:52] ppisati: yep, it just adds a boilerplate and insertchanges updates it after you commit your changes [16:53] so... [16:53] the process is [16:53] startnewlrelease [16:53] debian/rules insertchanges [16:53] and then a commit UBUNTU: Ubuntu-2.6.X-Y.Z [16:56] yep, from what I read from KernelMaintenance page in wiki [16:56] ah, the wiki... :) [16:56] ppisati: https://wiki.ubuntu.com/KernelTeam/KernelMaintenance [17:09] but it seems to me we should always use maint-startnewrelease instead of startnewrelease rule, correct? (at least that is the procedure on stable proposed releases) [17:10] herton: yep [17:10] herton: maint-startnewrelease [17:10] herton: apply stuff [17:10] herton: and then close it [17:10] i will tatoo these steps on my right arm so i won't forget it anymore [17:11] hehe [17:11] * herton --> lunch [17:11] herton, ppisati maint-startrelease also gets you the abi files and applies them to the commit [17:27] sorry, but still can't get my changelog updated [17:30] e.g. http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-lucid.git;a=commit;h=538abd32e3b8b1fb152927c9dfbd7b43f301e377 [17:30] i miss the modification to changelog that this commit has [17:30] i know, i did it in the past [17:30] but i don't remeber how i got the stuff there [17:31] is there some magic script? or what else? [17:31] ppisati, i can walk you through it [17:32] bjf: ok [17:32] i know i did it [17:32] ppisati, did the "maint-startnewrelease" do the right things? [17:32] yep [17:32] ppisati, thats the fdr insertchanges t [17:32] did you do a "fdr insterchanges" ? [17:32] that makes those changes [17:32] - CHANGELOG: Do not edit directly. Autogenerated at release. [17:32] - CHANGELOG: Use the printchanges target to see the curent changes. [17:32] - CHANGELOG: Use the insertchanges target to create the final log. [17:33] yes [17:33] the key lines are those and need to be there [17:33] that's what i get [17:33] also does fdr printchanges print anything [17:33] nothing [17:33] empty [17:33] as that is what gets inserted instread [17:33] printchanges i mean [17:33] ok then that likely means your previous commit has the wrong format [17:33] printchanges looks backwards for the UBUNTU: Ubuntu-xxxx line [17:33] and if thats not formatted right in the previous version it breaks [17:34] the tree is this one: [17:34] http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-lucid.git;a=shortlog;h=refs/heads/fsl-imx51 [17:34] commit 861bf6eb84e2b72966a881ce523661ae850b43ef [17:34] Author: Tim Gardner [17:34] Date: Thu Mar 10 01:57:49 2011 -0700 [17:34] Ubuntu-2.6.31-608.25 [17:34] and indeed its wrong [17:34] shall i... amend it? [17:34] it is missing the UUBNTU: prefix [17:35] i would ammend it yes [17:35] ok [17:35] let me try... [17:35] bjf, you have a script for checking a tree tip has been correctly closed i think? [17:36] apw, yes, "verify-release-ready" [17:36] ppisati, ^^ that is a good thing to use [17:36] ok [17:37] any way to amend the -16 commit? [17:37] w [17:37] ppisati, which one ? [17:37] the broken one [17:37] oh you mean the one which is 16 back from HEAD [17:37] yes [17:37] git rebase -i HEAD~16 [17:37] then change the first line to edit [17:38] then git commit --ame [17:38] then git rebase --cont [17:38] * ppisati tries... [17:42] ppisati, is that commit on master branch and was it in a release? [17:42] sconklin, Hey Steve, you still need me to write a request to revert that max_pfn_mapped patch for Lucid and Maverick or will it be done just based on the failure [17:43] bjf, its on an fsl-imx51 branch [17:43] apw, yes, but if that commit was part of a release, that history shouldn't be re-written [17:43] smb I already reverted it for both [17:43] sconklin, Ah ok good [17:43] bjf, that specific commit is safe as it is now in a tag [17:44] ok [17:44] and all of our arm kernels are rebase anyhow [17:44] so that having a non-linearaity isn't a big deal [17:47] everytime i close a release [17:48] i need a tracking bug? [17:48] i mean [17:48] since there's a big pile of cve [17:48] i wanyed to do it in steps [17:48] so, now i'm doing this [17:48] with some cves [17:48] and i'll ask for a pull [17:48] but i don't want an upload [17:49] well we don't have to close the release to do a pull of it [17:49] cool :) [17:49] so just do the startnewrelease, add gthe patches [17:49] and then just ask for a pull [17:49] ok [17:49] just leave it open [17:49] k [17:49] so, everytime we close, we upload and thus i need a tracking bug [17:50] yeah thats a pretty accurate [17:50] ppisati, https://wiki.ubuntu.com/KernelTeam/StableHandbook/StablePatchApplication [17:50] ppisati, that explains all the steps [17:50] ppisati, but for the meantime we can leave it not closed, [17:50] we often don't have it closed in the repo [17:50] with things sitting there waiting for more before the release gets closed [17:51] ok, so let's see it the other way around [17:51] i have a backlog of... [17:51] 50? 60? patches waiting to be applied [17:52] whats' the best policy? [17:52] well you are applying about 15 a day [17:52] s/policy/strategy/ [17:52] yes [17:52] i would apply them as i go, pushing the branch to zinc in my public repo ready for pulling [17:52] then say weekly ask for it to be pulled [17:52] when _all_ are applied we can close it and give it stable for handling out [17:53] so the there will be just one big test session at the end [17:53] yeah most of these cves are 1) generic and 2) already tested [17:53] so the chance of regression for any which are just cherry-picks is pretty low [17:53] ok [17:53] deal [17:53] i would test them at the end and cross my fingers :) [17:54] actually i don't have the hw [17:54] so from time to time i compile a kernel and "ship" it to someone to do, at least, a boot [17:54] i would probabally produce one weekly as i went and ask #arm to test it [17:54] k [17:54] as it takes hours to build :( [17:55] well [17:55] i do it in about 15/20 minutes [17:55] i cross compile [17:55] all the other arm guys do native compilation [17:55] but it's really too slow [17:55] * ppisati start a new recompilation... === ericm_ is now known as ericm-Zzz [18:06] * apw wanders off to the pub [18:19] * smb wanders off into the weekend [18:33] <-jealous :) [20:34] apw: so, what happened with CVE-2010-3880 and maverick, exactly? [20:34] kees: net/ipv4/inet_diag.c in the Linux kernel before 2.6.37-rc2 does not properly audit INET_DIAG bytecode, which allows local users to cause a denial of service (kernel infinite loop) via crafted INET_DIAG_REQ_BYTECODE instructions in a netlink message that contains multiple attribute elements, as demonstrated by INET_DIAG_BC_JMP instructions. (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2010-3880) [20:44] kees, it's in the changelog as having been fixed, will look at the commit [20:46] kees, the commit is there, it points at bug 711865 [20:46] Launchpad bug 711865 in linux-fsl-imx51 "CVE-2010-3880" [Undecided,In progress] https://launchpad.net/bugs/711865 [20:46] kees, do you believe there is an issue with the commit? [20:57] bjf: I don't know, apw's script flipped it from released to "pending" with no version [21:02] kees, _that_ i have no idea about [21:07] bjf: yeah, I'm scratchin' my head too :) it was the only thing I saw in the latest merge that looked funny === JayFo is now known as JFo [23:29] Um, hi there. I'm running Kubuntu Natty and I'm experiencing strange freezees (seemingly at random but mostly occur when playing a video). The video stops, I can kill the player, but if I try to type someting in for ex. konsole it doesn't respond, then everything stops responding (can't run anything, can't close anything) so I can only do SysRq + RSEIUB -> http://paste.ubuntu.com/613947 <- kernel log from before the freeze, what is [23:29] exatcly going wrong there?