[04:34] how does one get bug fixes into the official kernel? [04:35] same as any other package, put a patch in the bug [04:37] ok well this is the bug I am talking about [04:37] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/278648 [04:37] Malone bug 278648 in linux-ubuntu-modules-2.6.24 "[regression]snd-hda-intel sound input does not work at all with Conexant CX20549 (Venice) chips " [Undecided,New] [04:38] it's been almost a year since a fixed has been out ... and now it's come to updating to a kernel from a ppa repo ... [04:38] so what else needs to be done here? [04:39] crimsun: ^ you are the assignee [04:45] not talking to me ... right!? [04:47] zetheroo: thats right. Asking the bug assignee to comment. [04:47] lifeless: oh ok ;) [04:59] so there is nothing really that I can do ... [05:28] zetheroo: most probably that needs to get into 2.6.32.y before it can get into Lucid's kernel [05:28] zetheroo: foremost, however, it needs to make it into upstream sound-2.6 master [05:29] zetheroo: I've been extremely busy with other bugs, but I will look at it this week. [05:29] ok, thanks so much [05:30] anything I can do? [05:30] I am no programmer :P [05:30] just a humble user ... [05:56] crimsun: are you employed by Canonical ? [05:56] hope that is not a rude question [06:25] is there any way pmu_battery module could be loaded on PPC by default? Without this module the power management does not work at all. === mcasadevall is now known as Guest20793 [14:24] JFo, are you aware of any reports of Lucid LBM wireless drivers being unable to read firmware? [14:25] not off the top of my head tgardner, but I can look [14:25] JFo, 'preciate it [14:26] my pleasure [14:27] apw: I've got some questions on how the mainline kernel is built [14:27] you got a few mins? [14:28] cnd, maybe by thursday. [14:28] heh [14:28] cnd wassup :) [14:29] tgardner, If its atheros there was some time with problems. But lucid should be past that [14:29] apw, I tried using make-kpkg to make a custom kernel package based on mainline git tree [14:29] does make-kpkg work on any of our packages? [14:29] smb, iwlagn LBM refuses to read firmware. [14:29] I found that it mostly worked except that the debian control scripts that come with make-kpkg are old and don't work right [14:30] so a kernel from kpkg won't generate an initrd image [14:30] tgardner, Thats something different then I guess [14:30] i thought debian packages were supposed to contain them and indeed ours do? [14:30] apw: so, I was wondering how your mainline builds are built [14:30] apw, the initrd is generated by the postinst script [14:30] smb, yep, that something different, though I suspect its a generic problem. some interaction with udev but I haven't quite pinned it down. [14:31] tgardner, is it in linux-backports-modules-2.6.31? or later than that? [14:31] JFo, for Lucid [14:31] ah, I see [14:31] one sec [14:31] cnd they are built by checking out the commit of interest, then checking out debian* from the tip of the release you want them built for, then rebuilding the configs and building [14:32] tgardner, Well the intermediate problem on atheros was the added EEPROM checking code which was buggy for a certeain time [14:32] tgardner, We still need to get the new LBM up for karmic to fix that [14:32] smb, nah, this is a straight up failure after calling udev/firmware. [14:32] apw: how is versioning done, since the tip of the current release has the changelog for that release? [14:33] apw: or, where's the scripts that manage all this, I could just poke around myself [14:33] versioning is overridden with somethign c-o-d specific. -999 abi and the date/time as upload number [14:33] cnd, scripts are in the kernel-tools repo, in the mainline-builds directory [14:33] ahhh [14:34] tgardner, I only see 5 bugs for .32 http://bit.ly/cN10Nh [14:34] we still need to figure out what to do with kernel-package though... I'll try to figure something out [14:34] but there is the issue of whether there are some hiding in the 10k [14:34] that weren't reported against LBM [14:35] JFo, yeah, but it looks like nobody has encountered this firmware issue. [14:36] right [14:37] only bug I have found in the general population was a firmware filename issue [14:41] cking: will arrandale be supported in karmic at any point? or just lucid forward? [14:41] cking: meaning full platform support (audio, usb, processor, graphics, etc.) [14:44] apw, Just FYI, todays c-o-d mainline kernel explodes into my face. Though thats somewhat expected in the merge window [14:45] smb, completely expected :) [14:45] apw, A pity as the kvm patches got merged there. :-P [14:46] cnd, I'm only focusing on the low-level issues with arrandale in Karmic at the moment (e.g. HPET, TLB workarounds, TSC warping etc) [14:47] cking: I've got a few people who opened bugs about platform issues with arrandale on friday [14:48] cnd, that don't surprise me [14:48] cking: so there will be support, but not ready yet? [14:50] smb, ahh :/ [14:56] apw, I'm pushing stuff to your LBM repo, added a preempt flavor. -mta is gonna need some love as well. [14:56] -meta* [14:56] tgardner, ok ... i can handle -meta [14:57] so perhaps i'll aim for for starting the process first thing tommorrow [14:57] apw, I'm seeing a lot a -prepare noise in LBM so I'm gonna have a look at that as a possibility for this firmware issue [14:59] tgardner, ack [15:01] tgardner, i pulled in your two previous updates and released them so we could get on testing the drm update [15:01] i am expecting a respin of everything anyhow for tomorrow [15:03] apw, saw that [15:07] tgardner, so if i get you a test image for the qcm-msm i assume you can test it? [15:07] apw, no hardware yet. [15:08] tgardner, oh ... hrm [15:08] so i'll just rebase it, and build it, and call it good [15:08] apw, yep [15:19] apw, red herring in LBM. all the build noise is from ALSA [15:27] cnd: thats the intended behavior of kernel-package 12.x unfortunately :( [15:28] Sarvatt: why is that? [15:28] you have to manually add the hook scripts to /etc/kernel [15:28] Sarvatt: what if we put the ubuntu control scripts into the ubuntu kernel-package package? [15:29] /usr/share/doc/kernel-package/README.gz explains it [15:30] i think its a *horrible* choice but the author wanted it that way, i still havent got it fully set up right and i just build the initrd and update-grub manually after install [15:31] Sarvatt: that's a huge readme, can you just explain what you are meaning? [15:31] Sarvatt: I've gotten make-kpkg to autogenerate an initrd for me [15:31] the problem was the included debian/postinst doesn't work in ubuntu [15:34] oh really? [15:34] /usr/share/kernel-package/examples/etc/kernel/postinst.d/initramfs [15:34] that doesn't work? [15:34] Sarvatt: I didn't see that file, but I'll take a look [15:35] Sarvatt: I bungled up the first attempt I made at the correct debian/ scripts [15:35] cp /usr/share/kernel-package/examples/etc/kernel/postinst.d/initramfs \ [15:35] /etc/kernel/postinst.d/ [15:35] cp /usr/share/kernel-package/examples/etc/kernel/postrm.d/initramfs \ [15:35] /etc/kernel/postrm.d/ [15:35] so I'm rebuilding [15:35] Sorry for the long paste there [15:35] we'll see how it turns out [15:35] then in /etc/kernel-img.conf you add [15:35] postinst_hook = update-grub [15:35] postrm_hook = update-grub [15:36] Sarvatt: so there's some hooks that are added to run in the postinst script? [15:36] i haven't tried that yet, just found this readme yesterday but i've had this problem with kernel-package since 12.x came out over a year ago :( [15:37] yeah all the hooks are now in /usr/share/kernel-package/examples and you have to manually copy them to /etc/kernel/whatever for them to be added during make-kpkg [15:37] Sarvatt: ok, so that's the upstream kernel-package way it seems [15:38] I want to make a mainline kernel build that mimicks the ubuntu kernel as much as possible [15:38] for that to work, I think I need to just grab the ubuntu kernel's control scripts [15:38] and override the make-kpkg control scripst [15:39] so perhaps we could ship the ubuntu control scripts as part of the kernel-package package [15:39] not enabled by default [15:39] but still available for you to use [16:08] doing it the mainline kernel way uses a totally different workflow than make-kpkg would since it requires the ubuntu git repos being around, why dont you just use the kernel-tools stuff instead of kernel-package? [16:11] I use git pull && cp /boot/config-whatever .config && make oldconfig && make-kpkg clean && CONCURRENCY_LEVEL=3 INSTALL_MOD_STRIP=1 fakeroot make-kpkg --initrd -append-to-version=-sarvatt kernel_image kernel_headers >buildlog.txt 2>&1 && cp ../*.deb ~/downloads/current/ && cp buildlog.txt ~/downloads/current/ [16:11] straight from an upstream linux-2.6 directory [16:20] Sarvatt: the reason for what I'm doing is to try to do a git bisect between .32 and .33 [16:21] probably doing a real make-kpkg like upstream would work fine [16:21] but I want as few changes from ubuntu as possible [16:21] so I want to use the ubuntu .config [16:21] and packaging scripts [16:21] actually just using the mainline daily kernels is a good way to bisect without manually compiling [16:22] just start the bisect in your local tree, and find the date of each recommended commit and install the mainline daily from that day [16:22] Sarvatt: that's good point... I'm not sure why I didn't think of that earlier :) [16:23] though either way, figuring out how to make an "ubuntu-like" kernel using make-kpkg and documenting it is useful :) [16:28] smb, your ACK to the touchpad patch was understandably done with much consternation and gnashing of teeth === bjf is now known as bjf-afk [16:30] cking, Its shocking how much bad hw is out there [16:30] it's shocking how much of it comes our way ;-) [16:34] cking, Or only we care to do it automatically. I guess it could be worked around by a module parm. Or is it only possible that way? [16:34] cking, I know there is a i8042.noloop... [16:35] I don't believe it is possible to work around it - an there are a shed load of broken bits of hardware coming out nowadays [16:37] * smb wonders how many millicents you safe by not implementing the auxloop command [16:40] * cking wonders if other OSes check for AUX LOOP [16:40] Hrm, looks likle the i8042.noloop may be a workaround - but it's hard for the user with a default install to figure that is needed when their mouse doesn't work [16:42] Is it a known problem that when my laptop comes back from suspend, Xorg crashes? [16:43] and I get the "Your X is in low res mode" [16:49] BenC, what laptop is it ? [16:49] nvidia intel graphics ? [16:49] dell latitude 420, intel [16:49] D420 [16:50] not known then [16:50] latest lucid [16:50] my D420 works ok [16:51] easily reproducible, so I can get more info [16:51] some nvidia users were seeing the same thing [16:51] started this past week [16:52] I was going to take a look at it myself cause I thought it may be a driver issue [16:55] it's been happening for me for longer than a week [16:55] maybe 3 or so [16:56] BenC, got a bug report? [16:56] not yet [17:24] JFo: call me back? [17:45] JFo: did you figure out why 531511 isn't showing on the list? I think you need to add "linux-backports-modules-2.6.31" to the longnames list in KernelBugList.py [17:46] JFo: I can send you a quick patch if that'll help [17:46] yeah, I think that is the reason the numbers haven't been adding up [17:46] if you don't mind [17:46] JFo: sure, just a sec [17:47] thank you === bjf-afk is now known as bjf [19:04] cking, apw: that mem=nopentium fix, is that enabled automatically in lucid kernels? [19:05] no [19:06] cking, yes [19:06] commit 97684193e8c6fea095adcc4928f389b090559fc6 [19:06] Author: Colin Ian King [19:06] Date: Fri Feb 19 15:16:35 2010 +0000 [19:06] UBUNTU: Disable 4MB page tables for Atom, work around errata AAE44 [19:06] [19:06] BugLink: https://bugs.launchpad.net/bugs/523112 [19:06] Malone bug 523112 in linux "Intel Atom CPU can oops because of bug listed in Intel errata AAH41 and AAE44" [Medium,Fix committed] [19:08] apw, which kernel is that in? [19:08] I'm seeing some odd freezes on my new dell mini [19:10] -16 [19:11] oh yes, my slip [19:12] apw, ok I don't have anything to report on that kernel (yet...) [19:14] apw, you OK with me uploading Lucid LBM ? it'll fix firmware load problems that have existed for the last couple of revs [19:15] tgardner, sure [19:16] apw, this'll also have a preempt flavour (in case you've not done -meta yet) [19:16] i'll get meta sorted out now [19:16] let me know when its uploaded [19:16] is it pushed? [19:16] yep [19:16] gimme 10 minutes and I'll have it uploaded [19:19] tgardner: what was the fw issue? [19:20] cnd, new compat firmware load module plus new udev rules [19:21] apw, bombs away. lunchtime === Hedge|Hog is now known as Hedgehog === Hedgehog is now known as Guest40256 [20:10] apw, still around? [20:11] JFo: I know we're tagging bugs "needs-upstream-testing", I can't remember if we're then tagging ones that have been tested. [20:11] JFo: something like "tested-upstream" [20:11] I don't remember, but I think we have simply been removing the tag. I actually haven't been using it [20:11] I need to [20:12] JFo: yah, I'm wondering if you'd prefer we just remove the tag or tag them something like "tested-upstream" [20:12] oh, my preference would be 'tested-upstream [20:12] ' [20:12] so I can isolate those [20:12] programmatically [20:13] JFo: cool, I'll pass on the info [20:13] only thing is, we run the risk of having bugs with tons of tag clutter [20:13] cool [20:42] tgardner: I though karmic didn't require an sru for linux-firmware-nonfree because it's in multiverse? [20:43] linux-firmware may need an SRU for linux-firmware though... [20:43] for jaunty that is [20:44] cnd, ah, I can never remember what needs 'em. I'm sure Martin will tell you [20:44] tgardner: so the jaunty and karmic stuff has been uploaded where? to -proposed? [20:44] I don't see them on https://launchpad.net/ubuntu/+source/linux-firmware-nonfree [20:45] cnd, they're being held until released by an archive admin. should happen by tomorrow... [20:46] tgardner: ok. I'll add an SRU note in the bug too [20:46] thanks [20:46] tgardner: also, why did you need to change the versions? [20:46] for jaunty and karmic [20:46] I thought the packages had new version numbers [20:48] cnd, policy is to add a minor version for existing package within a pocket. a good example is linux-firmware which exists in both karmic and lucid. neither can have the same version number except for the initial sync. [20:49] tgardner: so that basically means that once a package has been released for a new release (i.e. lucid), all packages for the previous release have to increment the minor version? [20:49] cnd, correct [20:50] k [21:06] tgardner: I noticed that no one is subscribed to linux-firmware-nonfree bugs [21:06] should the c-k-t or kernel bugs teams be subscribed, like is done for linux-firmware [21:07] cnd, well, I'm not real interested in supporting linux-firmware-nonfree (for obvious reasons) [21:07] tgardner: ok [21:07] so, I _don't_ want to be subscribed. I'd rather have community take care of it. [21:08] I understand that sentiment :) [21:08] * cnd thinks we should get some mythbuntu people to take a look at it, since it's almost all dvb bugs [21:09] works for me. I'm happy to sponsor uploads, but I don't want to put anymore time into it then that [21:10] tgardner: did you mean to mark bug 446454 as fix released? [21:10] Malone bug 446454 in linux-firmware-nonfree "Move dib0700 1.20 firmware out of linux-firmware-nonfree" [Undecided,In progress] https://launchpad.net/bugs/446454 [21:10] it's still in progress by you [21:10] cnd, the launchpad bot will change its state when the archive admin releases it [21:11] hmmm... guess the package is taking a while to release then [21:11] lemme chack, I can't remember all of these versions [21:13] cnd, huh, dunno why its in progress. looks like its done to me. [21:13] yeah, that's why I thought something fishy was going on [21:13] lemme chanck the changelog [21:13] check* [21:14] hey all [21:14] I need to install an upstream kernel to test a bug [21:14] I was given http://kernel.ubuntu.com/~kernel-ppa/mainline/ as a link [21:15] is there a PPA line I should add to sources.list to test? [21:15] jono, nope. you gotta do it by hand. [21:15] dpkg -i ... [21:16] tgardner, just install the package? [21:16] cheers tgardner [21:16] I am going to see if I can provide any more info on https://bugs.launchpad.net/bugs/525902 [21:16] Malone bug 525902 in linux "Network card keeps dropping the connection" [Undecided,Confirmed] [21:17] jono, yep. we figured if you're not savvy enough to install a deb by hand, then you probably shouldn't be running mainline crack. [21:17] tgardner, wise [21:17] I assume I install -headers before -image? [21:18] jono, I just read the bug. Instead of mainline try linux-backports-compat-wireless [21:19] tgardner, is that in the Lucid archive? [21:19] jonhang on, I'll get the proper name [21:19] cheers [21:19] jono, linux-backports-modules-wireless-lucid-generic [21:20] but wait until tomorrow. I just uploaded a firmware fix today and its still building [21:20] tgardner, ahhh cool [21:20] will do [21:20] so are you aware of this bug? [21:20] jono, thats essentially 2.6.33 bits [21:20] cool [21:20] it just worried me that this could take down a bunch of T400s [21:21] jono, still reading the bug [21:21] ah, shirley-peak [21:21] ahhh [21:21] is there any chance of getting linux-backports-compat-wireless in the netboot img ? [21:22] so pgraner is talking to robbie about how to pursue this. its caught mark's interest as well [21:22] tgardner, this bug has? [21:22] not this bug per se, but the 5K intel series issues [21:22] tgardner, gotcha [21:23] well if I can help do any debugging, just give me a yell [21:23] jono, what kind of AP do you have, and is it set for 802.11n ? [21:23] tgardner, I have had this issue with multiple access points [21:23] jono, same room I assume? [21:26] tgardner, I have had the issue in the same room and when not in the same room [21:26] most typically the AP is in a different room [21:26] jono, as long as its pretty close it ought to work fine [21:26] tgardner, so is this bug an issue with range? [21:27] in Karmic I had exactly the same AP and network card and no bug [21:27] jono, I don't think so, it appears to be related to the 5K adapter. The 3K and 4K series seem to work fine. [21:28] jono, well, lets see if tomorrow's LBM makes any difference. [21:28] tgardner, no worries, I will test tomorrow and then update the bug [21:28] k, I'll get subscribed [21:29] thanks tgardner :) [22:03] crimsun: you about? [22:52] JFo: https://bugs.edge.launchpad.net/~kernel-bugs/+patches [22:53] bdmurray, excellent, thank you. This looks great === sconklin is now known as sconklin-away [22:56] JFo: well, thanks the launchpad team and bryceh - I was just pointing it out ;-) [22:56] heh [22:57] yep, thanks to them then, and thanks for pointing it out [23:03] I just wish that it gave a way to exclude the 'Fix Released' patches... [23:18] bryceh, yeah, I noticed... it is cool though. I can work with it [23:18] :)