=== amitk-afk is now known as amitk [08:31] Morning apw [08:37] lag y [08:37] lag, yeah morning, sorta [08:38] apw: Sorta? Still up from last night? [08:39] lag more sorta up, its been a long week [08:39] apw: You're still in bed? :) [08:39] having to get up at 8 on a sunday and go to work is not the best cure for tiredness [08:39] heh [08:40] Yeah, that sucks :) [08:40] How did it all go though? [08:40] Everyone looks happy enough in the photos [08:40] still i think it all went out near to hitch free, as near as sometihng that complex can ever be so i suspect [08:41] though trying to even sync an iso today is taking hours, so i assume we are being hammered to death still [08:41] That's a good thing :) [08:41] Mo' users [08:42] i am seeing about 25KB/s from here, and i can smell the DC from here [08:42] so i assume we are red-lining all our b/w [08:42] :) [08:42] So what now? [08:43] a lazy day for me for sure [08:45] And for the rest of us saps? [08:51] heh *crack* back to digging out the ore [08:52] this is the slack time, there is a lull between now and UDS where we start to think about what we are going to do [08:52] and we start the sprint again [08:53] Okay [08:53] I guess I'll just crack on with my bugs then === yofel_ is now known as yofel [09:29] where's the URL to the kernel related blueprints for natty? [09:37] cking_, apw was messing with that stuff earlier last week === smb` is now known as smb [09:39] cking_, they are likely all lost now as someone renamed them [09:39] let me find them [09:43] cking_, https://blueprints.edge.launchpad.net/sprints/uds-n?searchtext=kernel [09:43] apw, ta [09:43] cking_, all of them are on that page, many have hardware-prefix, a couple have other- [09:43] new fun track prefixes [09:43] that messes up all my links gah [09:47] is there any naming scheme being used - looks a little ad-hoc to me [09:50] cking_, yes - smb, what's our official "track name" then? [09:54] cking_, whatever they think could match. I don't think it was a single track that apw found [09:54] gah [09:55] cking_, yep the tracks are 'ideas' oriented this time not team oriented [09:56] so like cloud, hardware, etc [09:57] fwts falls into hardware I suppose, [09:57] i guess [10:00] smb, I see a lot of the blueprints named "hardware-kernel-n-some-title" - doesn't that conflict with - as you mentioned earlier. I'm confused [10:02] cking_, Are you always taking me up to the word? [10:02] It was more of an impression [10:03] I yet have to find my blueprints again to even know what they changed them into [10:03] cking_, There was a mail by jcastro where he was telling the scheme I think [10:04] ah, /me looks [10:17] how do I export a single commit with git format-patch? if I just do "git format-patch abcdef" it'll take everything from abcdef+1 to HEAD, which is not what I want. [10:18] diwic, with -1 [10:18] git format-patch -1 [10:19] smb, thanks. Strange, this was not documented on "git help format-patch" [10:21] diwic, Likely it is hidden somewhere in git-rev-parse in a very unovious way [10:21] diwic, I usually found a "ask apw" much more effective [10:22] diwic, git format-patch -1 [10:22] diwic, or git format-patch .. [10:23] (where A is before the range as normal) [10:24] apw, right, could have done git format-patch abcdef^..abcdef as well [10:24] yep [10:24] -N is easiest for small values of N [10:24] aha, so abcdef -2 will take abcdef and abcdef+1? [10:25] apw, when registering a blueprint I'm asked "For:" ..(The project for which this proposal is being made.". If I put in "ubuntu" LP complains that there are too many matches on this keyword. How did you fill in this field? [10:26] cking_, for as in project, thats ubuntu [10:29] cking_, pretty sure that i just typed it in and moved on [10:29] cking_, yeah just shove in the text ubuntu and move to the next field, the searchy thing blows a sprocket [10:29] argh, it's now lost it *again*. LP time outs are getting me crosser by the moment [10:30] Who will do the ti-omap4 packaging for natty? [11:02] Blueprint saga now complete· [11:08] cking_, how many am i looking for ? [11:09] apw, just one, I've been wrestling with other things too, it's been a background task for this morning === diwic is now known as diwic_afk === diwic_afk is now known as diwic === ivoks-afk is now known as ivoks [13:38] smb: $track-$team-n-$title is the format === BenC___ is now known as BenC [17:46] vanhoof: are you working today? [17:47] vanhoof: I assumed you'd be recovering from jet lag and observing the holiday anyways [17:50] ogasawara: Congratulations :) [17:50] lag: thanks :) [17:51] What's next? [17:52] lag: world domination [17:52] Sounds like you're up to it === xfaf is now known as zul [18:28] hi people [18:28] i wonder why on ubuntu/debian nouveau on recent (nvidia gt230m and 310m) cards i get a black screen on boot [18:29] while on fedora it works great since lots of months [18:29] it's a problem related to some of debian/ubuntu kernel patches? [18:31] ogasawara: kinda sorta, we're on holiday :) [18:31] ogasawara: wanted to get that tested though, gonna head out to lunch [18:32] vanhoof: am gonna build you another test kernel with those additional suspend related looking patches. But don't test it till tomorrow and just enjoy the rest of your day. [18:33] ogasawara: kk :) [18:33] bo it may be a firware issue, or lacking of patches ... we'd want a bug filed from the machine [18:33] ogasawara, morning ... [18:33] looks like its all gone out ok ... congrats on your first born [18:34] apw: heh, hopefully baby #2 will be just as painless [18:34] ogasawara, heh i bet you do :) [18:37] apw: now i'm on fedora, how can i report a bug in a machine with a black screen from the first second of boot? (but not freezed, in fact i heard the gdm sound) [18:37] bo, fair enough then === ppetraki is now known as ppetraki-away [18:38] on fedora-kernel they told me that they don't ship any nouveau code that isn't upstream [18:38] i don't know if it's a kernel problem or whatelse! [18:38] bo and what kernel version is fedora shipping [18:39] most likely a combination between that and the mesa [18:39] now, 2.6.35.6-39.fc14.x86_64... but it works with many past fedora kernel [18:41] there are a lot of forum topic on this problem, i tought it was a card problem, but some months ago i find that on fedora worked [18:41] bo, and you were testing lucid or maverick [18:42] didn't work on both [18:42] and also on debian alpha [18:44] whats the card bo ? [18:44] before maverick i used closed driver, and a put on xorg the EDID file description path, [18:44] and the it worked [18:44] so maybe it isn't related to nouveau [18:44] yet nouveau works on fedora [18:45] bo, what nvidia card ? [18:45] 01:00.0 VGA compatible controller: nVidia Corporation GT216 [GeForce GT 230M] (rev a2) (prog-if 00 [VGA controller]) [18:46] RAOF, Sarvatt, know anything about noveau not working with the above card on maverick ? [18:47] bo i assume your testing was on the live CD image as you are on fedora ? [18:47] no, now i installed fedora [18:48] cause on lucid i managed to have graphic with closed driver and EDID file on xorg [18:48] in maverick it doesn't work [18:49] bo, how long ago did you switch away, ie. when did you last test maverick [18:49] it works just with vesa [18:49] i tested maverick yesterday [18:50] i use fedora since august [18:50] yesterday i tried to come back on ubuntu [18:51] before these months I always used ubuntu o debian with closed driver [18:55] bo, thanks [18:55] now i remember that i can connect the notebook to my tv via hdmi, in this configuration i DO see things on video [18:55] and i can report bug [18:56] so you should be able to do that via the liveCD at least [18:56] yes [18:56] now it seem to me just a monitor problem [18:57] we may be able to tell that from the info in the bug [18:57] what kind of infos do you need? [18:57] i mean a log or whatelse? [18:57] perhaps fedora has edid hacks, which would be common to all drivers and not specific to edid [18:58] bo id you file a bug against the xorg driver nouveau we should get the kernel drm information too [18:59] so i'll ask to fedora team i they've done an hack! [18:59] sorry, how can i see kernel drm info? kernel.log? [19:02] /sys/kernel/debug/dri [19:02] ok, thanks, i have to go now, see you later [19:03] and /sys/class/drm/card0-* [19:03] (i'll read later) === ivoks is now known as ivoks-afk [19:25] abogani: What is your gpg key ID? [19:25] abogani: nm === ppetraki-away is now known as ppetraki [20:01] ogasawara: i know i shouldn't be working, but i couldnt resist :) [20:01] ogasawara: success w/ v2 [20:01] * vanhoof updates thread [20:01] vanhoof: sweet, I'll build one more to narrow it down between the two [20:02] vanhoof: but seriously, don't feel the need to test it till tomorrow [20:02] ogasawara: you keep sending them, i'll keep testing :) [20:02] ogasawara: im just ignoring everything else in my inbox atm ;D [20:03] but i do feel its time for a nap lol === ivoks-afk is now known as ivoks === ivoks is now known as ivoks-afk [21:17] Hi, I need some debugging help [21:18] Maverick kernels occasionally seem to malfunction on my box, manifesting as system monitor graphs showing the CPUs very busy with system time [21:18] top shows nothing - how can I trace which kernel task is to blame? [21:31] 10.04 appears to have KSM enabled, but I can't get madvise() to return success on when I pass the MADV_MERGEABLE flag [21:31] anyone else use KSM successfully? [21:36] AFAIK kvm uses it and I seem to remember seeing over 500MB of merged pages on some boxes (running libvirt with kvm) [21:37] freetz: What error do you get? [21:37] einval [21:37] which should only be returned if ksm is disabled [21:37] freetz: Stock maverick kernel? [21:37] freetz: Oh, sorry, 10.04, you say. [21:38] root@chopin:~# cat /sys/kernel/mm/ksm/pages_shared [21:38] 136582 [21:40] (10.04.1 on 2.6.32-24-generic 64bit) [21:42] yes [21:43] grep KSM /boot/config-2.6.32-25-generic [21:43] returns CONFIG_KSM=y [21:43] i get that too for pages_shared [21:43] i enable ksmd by setting /sys/kernel/mm/ksm/run = 1 [21:44] madvise will also return EINVAL if your addr arg isn't page-aligned. [21:45] hmm, i'll check on that real quick [21:45] ogasawara: v4 ftw [21:45] * vanhoof grabs a beer [21:46] freetz: That's really (AFAICT) all it can be. madvise doesn't give EINVAL if you don't have ksm enabled, only if you haven't compiled it in at all. [21:46] isn't valloc deprecated? [21:46] or is there a better way to ask for page aligned memory [21:47] vanhoof: \o/ [21:47] memalign [21:47] freetz: ^ [21:47] woop just saw that... [21:49] QEMu (which I /know/ succesfully uses KSM) uses posix_memalign to ensure properly aligned pages. [21:49] fwiw.