=== bjf is now known as bjf[afk] [00:24] squarebracket: which release? Also, what is missing from the currently provided driver that would require it be repackaged in lbm? [00:25] the latest kernel module (which supports more/newer devices) [00:26] ogasawara, what do you mean which release? like, lucid, maverick, etc? [00:26] squarebracket: right [00:27] ogasawara, well 2.6.35 has the latest code, so maverick doesn't need it yet (but it will, i'm sure) [00:27] everything earlier does [00:27] squarebracket: it's best to open a bug then and provide the justification for needing it. [00:27] ogasawara, there is one [00:28] ogasawara, let me find, one sec. [00:28] https://bugs.launchpad.net/ubuntu/+source/linux-backports-modules-2.6.35/+bug/606278 [00:28] Ubuntu bug 606278 in linux-backports-modules-2.6.35 (Ubuntu) "Add linux-backports-modules-input (affects: 6) (dups: 1) (heat: 218)" [Undecided,New] [00:29] oops, didn't know that would trigger a bot [00:30] ogasawara, without providing a lbm module, one has to compile the kernel module from the linuxwacom source (i was doing this for a while) [00:31] ogasawara, each and every there is a kernel update, i mean. [00:33] squarebracket: There's a dkms-based source package on revu which I shall look at today. That may be better than an l-b-m, particularly since it's not actually a backport. [00:36] RAOF, yeah, i know of said package. i believe it builds from the source on the linuxwacom sf site. someone suggested an lbm, which kinda made sense, but i can't say i actually know what constitutes a backport [00:38] RAOF, really, anything will work, so long as the end user doesn't notice :) thanks for taking a look [00:44] ogasawara, thank you :) [08:50] * apw yawns [08:52] apw: Have you seen the async-setup-outputs intel patch on dri-devel? It might be interesting if we want to shave off more boot time. [08:54] RAOF, as in each one in parallel? [08:54] (no not seen it btw, and yes it sounds intresting) [08:55] http://article.gmane.org/gmane.comp.video.dri.devel/48689 [08:55] As in: move all the probing, output setup, and fb creation to a thread, and not block the boot doing it. [08:57] Hm, I wonder whether this might cause problems when starting slpash screens while the drm fb might be not ready... [08:57] RAOF, ok, if you look at the patch it simply moves the connector setup async en-block, as a single threaded init [08:58] so in fact its as async as it is now, as we load it as a module which is probed async anyhow by udev [08:58] smb, and as its async already we already have the issue, which is why i had to patch all that crap in fb and drm [08:58] Ah, the magical intricacies of the kernel. [08:59] thats only useful if the driver is built-in [08:59] now if it probed all the connectors in parallel that might make sense either way, but its too small to be changing that [09:01] apw, sort of similar, though I think in your case there is something already there or half there and causes problems. On some slower machines I already saw cases where drm fb is not even half there. Causing vga16fb to be in use one the way up and drm fb on the way down. Not as annoying and critical as the stuff you had to cope with but strange to a user [09:02] smb, cirtanly it is easy to end up initialising vesafb, and then drmfb, and we don't handle that well at all right now at the plymouth level [09:03] Right, but it only looks odd but goes on and does not crash or hang the whole process as in the other case. === cking__ is now known as cking1 [09:05] cking1, If you want to shout at us you need to turn up the volume. ) [09:05] just done it [10:05] cnd, you are a per-package uploader for firmware right ... that should mean you are already an unbuntu member ... if not its broke, you shouldn't have to apply [10:06] cnd, you can tell if you go to this page: https://edge.launchpad.net/~ubuntumembers [10:06] for me it shows my membership via my upload rights [10:07] apw, nope [10:07] according to sabdfl, ppu doesn't make you a member :) [10:08] hrm, it makes you an ubuntu developer, which should as far as i know [10:08] and I'm not a member of the launchpad teams that I should be if I were a member [10:08] me feeling is that that is broken, i would cirtianly ask someone like cjwatson if its true [10:08] not a bad idea [10:08] I'll ping him [10:08] as you have interviewed enough to get upload rights but not enough to be a member, seems wrong, and contrary to what i was told [10:10] yeah, it's just one of those things where you can't tell sabdfl he's wrong when he's two feet in front you :) [10:10] even if that's the case [10:11] which is a bad idea when you don't really know [10:17] cnd, indeed not :) ... cjwatson or perhaps pitti would know for sure [10:17] lag, you need to fill out your application ... its all empty still, bad lag [10:17] apw: I'm not ready [10:18] I've only been here 3.5 months :( [10:18] 'course you are, just fill it in ... most of its boiler plate [10:18] Okay, when I have a moment [10:18] it'll be 6 months before long, and a month before the end you need it on the list so you get done at 6m, and ... you made me put something in ... so you have to :) [10:20] lag but my advice overall for any of this is to keep that top level page up every couple of weeks [10:20] as when you come to fill in what you've done you can never remember what you did [10:21] You only just did that! [10:22] :) [10:43] apw: A little better? [10:44] a little i guess ... [10:44] * apw sulks [10:49] smb: does the writeback regression also affect nfs? certainly seems so here [10:50] tjaalton, Could well be [10:50] i've got 100 nfs4 mounts and unmounting them takes several minutes [10:51] iirc .35 felt better [10:51] If you are able to experiment, there are test kernels referenced in the bug report. Only bad thing there is that I cheated a bit with abi numbers. So server should be ok, but desktops don't work that well [10:52] But yes, the symptom would be what you describe [10:52] And as it is a generic problem on writeback every fs hits the same issue [10:52] ok, I might be able to test it later [10:53] Maybe this also gets accepted for the next larger proposed upload. So it would show up in pre-proposed [10:53] soonish [10:54] /* Ordinary console input and output */ [10:54] fd = open (CONSOLE, O_RDWR | O_NOCTTY); [10:54] if (fd < 0) [10:54] nih_return_system_error (-1); [10:54] not a very elaborate error handling [10:55] given there are a few conditions that are retryable and should be [10:55] now that nfs4 shares are umounted, dpkg unpack is a lot faster [10:55] /* Copy to standard output and standard error */ [10:55] while (dup (fd) < 2) [10:55] ; [10:55] what if dup fails [10:55] owch [10:56] dup a while... dup forever... === yofel_ is now known as yofel [11:14] * No longer holds /dev/console open, so the SAK SysRq key will not [11:14] kill Upstart. (Bug: #486005) [11:17] apw, according to: https://wiki.ubuntu.com/UbuntuDevelopers [11:18] PPUs "are implicitly considered Ubuntu Members" [11:18] so it sounds like I don't need to go through the process [11:18] I'm still gathering info in #ubuntu-devel though [11:32] cnd, talk to whoever added your rights, likely cjwatson, he probabally missed something [12:21] hi [12:24] since i have upgraded to maverick with the .6.35 (2.6.35-14) I am facing serious problems with gfx card. how to detect if this is caused by KMS bug ? [12:24] I have no more gfx, and when booting to .6.32 kernel, I have gfx (X running) but without mouse/touchpad input [12:25] so it seems to me something caused by a kernel kms bug, I have ATI 3200 [12:54] I have this error when booting http://pastie.org/1083660 [12:55] kernel .6.35-14 ATI 3200 card, [13:12] boulabiar, and does it affect operations? [13:36] apw, I have no more graphics [13:55] boulabiar, if you have previous kernels please boot back into those and confirm exactly where this regression occurs [13:56] boulabiar, and file a bug with ubuntu-bug linux and include all that information [13:56] apw, I have only tested with 2.6.32 [13:56] I got X running, but with only keyboard input [13:57] apw, is there some ppa for precompiled kernels to not lose time compiling them ? [13:59] the previous kernels are all available from launchpad [13:59] https://www.launchpad.net/ubuntu/+source/linux [14:21] apw, if you speak about the 2.6.32 of Lucid, then graphics works with it, I'll test something else [14:22] boulabiar, ok so lucid kernle is ok, if you could try the older maverick 2.6.35 kernels that would be instructive [14:29] * apw goes to investigate some new kit [14:51] there was a way to set a timer to wake up a system from suspend, can anyone help me out where to echo it? [14:55] TeTeT: /sys/class/rtc/rtc0/wakealarm [14:55] TeTeT: download cking's scripts git tree to get a nice shell script to do this [14:56] TeTeT, http://kernel.ubuntu.com/git?p=cking/scripts/.git;a=commit;h=3577ea60c525f4e3d91b2f7afa009fa68e4a6ff8 [14:58] cking_ + amitk : thanks, glad I don't need to re-invent the wheel :) [14:58] :-) [15:12] weird, the system does not resume completely, it half wakes up, disk is spinning up, but screen stays black, then goes to sleep again === amitk is now known as amitk-afk [16:20] ## [16:20] ## Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting [16:20] ## agenda: https://wiki.ubuntu.com/KernelTeam/Meeting [16:20] ## [16:32] bjf, One thing I forgot over all the other stuff, is there a current branch for the mvl-dove pull request in Lucid that i can review? [16:32] smb, yes, there is one in my tree on zinc [16:32] smb, let me dbl check the branch name for that [16:32] Ok, thanks, then I have a look there in a min [16:33] smb, it is mvl-dove-5.3.2 [16:33] ack [16:34] apw: thanks for helping out with bug 554172...it has a large peanut gallery of subscribers :/ [16:35] Launchpad bug 554172 in linux (Ubuntu Maverick) (and 2 other projects) "system services not starting at boot (affects: 187) (dups: 15) (heat: 824)" [High,Confirmed] https://launchpad.net/bugs/554172 [16:58] ## [16:58] ## Kernel team meeting in one hour [16:58] ## [17:21] jjohansen, will you be attending the meeting today? [17:22] bjf: no, I'm working on a status email for you [17:23] jjohansen, cool, thanks === maks is now known as maks_ [17:38] what's ubuntu's position for xen? can someone have a look at the sponsor request for xen (bug #378240)? [17:38] Launchpad bug 378240 in xen-3.3 (Ubuntu) "Please merge xen-3.4 (3.4.0-2) from debian unstable (affects: 3) (heat: 33)" [Wishlist,Confirmed] https://launchpad.net/bugs/378240 [18:00] ## [18:00] ## Meeting starting now - #ubuntu-meeting [18:00] ## agenda: https://wiki.ubuntu.com/KernelTeam/Meeting [18:00] ## [18:20] calendar item should be visible now [18:21] apw, want to room for UDS or are you tired of me? :) [18:21] JFo, works for me [18:21] cool [18:21] I'll drop the details here then https://wiki.canonical.com/UbuntuPlatform/UDS/N [18:22] Reminds me to add things as well. cking_ you are up for the usual teaming, too? [18:23] smb, indeedy, works well for me === bjf changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Maverick Kernel Version: 2.6.35 || Ubuntu Kernel Team Meeting - August-17 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! [20:05] * apw gives up for the day [21:44] tgardner: ping [21:46] pgraner, yo [21:56] hello [21:56] I'm trying to rebuild one of the kernel packages as a first step in building a patched kernel package [21:56] can anyone help? [21:57] tgardner: Are you about? [21:57] lag, yep [21:58] tgardner: Any chance in having update-initramfs installed on the Orange please? [21:58] tgardner: Arm chroot [21:59] lag, in the ARM schroot, right? [21:59] I tried make-kpkg --initrd ... kernel_image kernel_headers and ended up with .debs, but when I installed them I couldn't boot [21:59] tgardner: Yep [21:59] so I'm confused about what I did differently from how the linux-image packages in apt are built [22:00] lag, lucid or maverick or both ? [22:00] I'll be using it on Maverick [22:00] tgardner: I guess it couldn't hurt to have it on both? [22:01] lag, yeah, I'll add it to the schroot scripts as well [22:01] tgardner: Would it be too much trouble to put it on Diamond and Roofer? [22:01] Will that add it to the other ones? [22:02] lag, that depends. its probably not a bad package to have in all lucid/maverick schroots [22:02] mdgeorge: looks like not much help here... which is a shame. [22:03] tgardner: Quite [22:03] doors: actually, I followed the link in the topic, and it looks promising [22:03] lag, hmm, its already installed in both schroots. what makes you think it wasn't? [22:04] tgardner: It's okay if we're testing random kernels, but I need to test them with the daily builds and hence need to make my own initramfs' [22:04] Really? [22:04] Hang on [22:04] lag, would I lie to you? [22:05] heh [22:05] I can't see it [22:05] update-alternatives update-language update-updmap [22:05] update-fmtutil update-language-dat [22:05] update-fontlang update-language-def [22:05] (maverick-armel)root@tangerine:/home/rtg# dpkg -l|grep initramfs-tools [22:05] ii initramfs-tools 0.97.2ubuntu1 tools for generating an initramfs [22:05] ii initramfs-tools-bin 0.97.2ubuntu1 binaries used by initramfs-tools [22:06] (maverick-armel)root@tangerine:/home/rtg# dpkg -S /usr/sbin/update-initramfs [22:06] initramfs-tools: /usr/sbin/update-initramfs [22:07] /usr/sbin isn't in my $PATH [22:07] Doh! [22:08] Is it in yours? [22:08] "which update-initramfs" ? [22:08] lag, normally, but remember that tangerine has a strange account setup. Maybe you need your .bashrc from zinc [22:09] I'll put it in manually [22:09] tgardner: Sorry to waste your time :) [22:09] lag, nah, this is really my fault. I should have corrected the account setup from the get go. [22:10] np [22:10] I'll sort it out tomorrow (it's late here) [22:10] lag, yeah, back to the pub .... [22:10] Quite :) [22:10] * lag_ leaves [22:24] yay conference wireless [22:29] Keybuk, heh it won't last ... never does [22:30] apw: if I keep believing that, it might [22:30] Keybuk, /me radiates positive thoughts towards the wireless [22:30] stop radiating! [22:31] heh [22:39] so, tangentically [22:39] I did some more digging [22:39] and it looks like there are situations where an open() can return -EIO ;-) [22:39] if you have a journalled filesystem [22:39] and there's a physical I/O error [22:39] so going into an infinite loop on EIO seems like a REALLY BAD IDEA