| lukehasnoname | I ran into a few bugs today, and I think they belong to either the installer or the kernel or the x team | 01:55 |
|---|---|---|
| lukehasnoname | I don't know | 01:55 |
| lukehasnoname | uvesafb and v86d errors on boot, and an error running post boot oem-config at boot' | 01:56 |
| lukehasnoname | Very obnoxious and crippling, unless you know what you are doing, which I only halfway do in this instance | 01:56 |
| lukehasnoname | http://mibbit.com/pb/3ui0Ts | 02:08 |
| lukehasnoname | I will likely mail the kernel and installer MLs after some things to do this evening. | 02:08 |
| lukehasnoname | I don't like being on Vista unless I'm gaming or .NET'ing. Help a bro out and figure this one out! | 02:08 |
| CarlFK | how did you install the driver ? | 02:13 |
| lukehasnoname | CarlFK: ... I just installed the OS, I didn't install anything manually. | 04:17 |
| lukehasnoname | Everything was done via the installer, which is why I didn't know if I should talk here (I assumed things that happened during boot would likely be involved with the kernel | 04:18 |
| lukehasnoname | ) or if I should put it to the installer guys | 04:18 |
| ripper | Hello there! | 04:43 |
| ripper | BenC, Hi Ben | 04:44 |
| === ripper is now known as AzzCoder | ||
| erichj | is bug #260675 going to be addressed in time for 8.10 release? | 06:47 |
| erichj | https://bugs.launchpad.net/ubuntu/intrepid/+source/wacom-tools/+bug/260675 | 06:47 |
| === mdz_ is now known as mdz | ||
| AnAnt | Hello, when will 2.6.27-6.9 be available in Intrepid ? | 09:32 |
| abogani | AnAnt: Now | 09:37 |
| AnAnt | abogani: thanks | 09:39 |
| AnAnt | abogani: will it enter the Queue, or will it immediately appear on the repos ? | 09:42 |
| abogani | AnAnt: I just reboot my laptop with it. | 09:43 |
| AnAnt | abogani: ? | 09:44 |
| AnAnt | abogani: rmadison still reports 2.6.27.6.7 for Intrepid | 09:44 |
| === abogani changed the topic of #ubuntu-kernel to: "Ubuntu kernel development discussion ONLY | Kernel Wiki: https://wiki.ubuntu.com/KernelTeam | Latest news: Moved to 2.6.27 kernel for Intrepid/8.10. | Kernel git trees: http://kernel.ubuntu.com/git | Latest kernel upload: 2.6.27-6.9 based on 2.6.27-rc9" | ||
| AnAnt | abogani: are you gettings updates from some PPA ? | 09:48 |
| abogani | AnAnt: No | 09:48 |
| AnAnt | -> http://eg.archive.ubuntu.com/ubuntu/pool/main/l/linux-meta/ latest is 2.6.27.6.7 | 09:49 |
| AnAnt | abogani: can you tell me which repo you got it from ? grep "intrepid.*main" /etc/apt/archives | 09:51 |
| abogani | AnAnt: http://paste.ubuntu.com/55212/ | 09:53 |
| AnAnt | ok, thanks | 09:59 |
| AnAnt | and sorry | 09:59 |
| === asac_ is now known as asac | ||
| === thegodfather is now known as fabbione | ||
| Keybuk | if I said "systemtap", would I hear screaming? :p | 13:16 |
| tseliot | Keybuk: hmm... maybe try with dtrace ? ;) | 13:17 |
| amitk | Keybuk: systemtap works? | 13:19 |
| ogra | amitk, oh, youre around ? | 13:21 |
| amitk | ogra: depends on what you want ;) | 13:22 |
| ogra | amitk, well, lool gave me tohe todo to look at the ath5k issues before release with you :) | 13:23 |
| amitk | ogra: ath5k issues as in why it doesn't work on the Q1 or to blacklist it? | 13:24 |
| ogra | he belives its Q1 specific that it doesnt work so it must be something with the HW combo | 13:24 |
| ogra | his hope is that we either geti it working or blacklist it inside the module against the device | 13:24 |
| ogra | (which i dont belive is possible) | 13:25 |
| amitk | ogra: it should be.. using a PCI device quirk | 13:25 |
| ogra | hmm | 13:26 |
| amitk | I can look at the Q1 + ath5k issues tomorrow. Did you try LRM? | 13:26 |
| ogra | works perfectly | 13:26 |
| ogra | the one issue i have is if i add a blacklist file to exclude ath5k, pm-utils autoloads it on resume, but i added a fix to pm-utils for that | 13:27 |
| ogra | everything else works fine | 13:27 |
| amitk | too bad... It would really be nice if we could get the open driver working. | 13:28 |
| ogra | yeah | 13:28 |
| ogra | well, we still have a week or so ..., so we could still try if you want to invest time | 13:28 |
| amitk | I'll look at it tomorrow. Does the -mid image use the lpiacompat kernel? | 13:28 |
| ogra | no, lpia | 13:29 |
| Keybuk | amitk: I haven't tried, people claim it does | 13:29 |
| rtg | amitk, ogra: we're going to end up backporting 2.6.28 wireless drivers before things start working well. A lot of stuff missed the 2.6.27 merge window. | 13:29 |
| ogra | they also have lrm, so we have a waorking solution in any case, no matter what we come up with | 13:29 |
| amitk | Keybuk: I tried real quick during the systemtap talk at plumbers conf. Didn't get it to work immediately... | 13:30 |
| amitk | rtg: is that the wireless work you mentioned you will be doing this week? | 13:30 |
| ogra | rtg, is that in intrepid timeframe ? | 13:31 |
| rtg | it'll be much like Hardy LBM is now. e.g., backported 2.6.27 drivers. | 13:31 |
| rtg | amitk: the wireless work I'm doing this week is just profiling the cards that I have to find out what works out of the box. | 13:32 |
| ogra | rtg, so i still need a workaround for the final image then (it would be bad if ubuntu-mobile wouldnt work on the reference platform we advertise) | 13:32 |
| rtg | ogra: what wireless card? | 13:33 |
| ogra | ath5k ... let me look up the exact chipset name | 13:33 |
| rtg | its probably AR2412 | 13:33 |
| ogra | AR242x 802.11abg | 13:34 |
| rtg | ah, AR2424. | 13:34 |
| ogra | on the samsung Q1 sepcifically, we have users claiming it works on a different HW combo with that chipset | 13:34 |
| rtg | ogra: I may have one of those. | 13:35 |
| ogra | (though unencrypted) | 13:35 |
| ogra | rtg, bug 274832 has some info | 13:35 |
| ogra | (note that it doesnt seem to matter if you are on -generic or -lpia, though the bug is about lpia here) | 13:36 |
| rtg | where is ubuntoid when you need it | 13:36 |
| ogra | https://bugs.launchpad.net/ubuntu/+source/mobile-meta/+bug/274832 | 13:36 |
| ogra | :) | 13:36 |
| rtg | ogra: there is an LRM package for lpia now, correct? Are you using madwifi or ath5k? | 13:38 |
| ogra | well, we have two images now, ubutntu-mobile and ubuntu-mid, mid uses lpia and mobile is just a desktop install with customized look and feel, both install lrm now and the wifi works with it, for both wifi doesnt work with ath5k | 13:40 |
| ogra | (on the Q1) | 13:40 |
| rtg | ogra: ok, I'm booting to see what Atheros card I have. | 13:41 |
| * Keybuk alive | 13:44 | |
| * Keybuk really alive | 13:44 | |
| Keybuk | amitk: linux-image-debug went away? | 13:46 |
| amitk | Keybuk: not that I remember. debian/control has an entry too | 13:51 |
| Keybuk | amitk: I can't see it in the archive | 13:52 |
| amitk | Keybuk: hmm.. there is no meta package for linux-debug though | 13:53 |
| rtg | ogra: all I have are ath9k devices. | 13:54 |
| ogra | meh | 13:54 |
| Keybuk | amitk: I can't see the real package, let along any meta packages ;) | 13:54 |
| amitk | Keybuk: yeah.. same here. Wading through git now... | 13:55 |
| ogra | linux-image-debug-386 | 13:57 |
| * ogra sees a i386 one | 13:57 | |
| ogra | ah, referring to linux-image-debug-2.6.25-2-386 | 13:59 |
| * amitk looks towards rtg and BenC | 14:03 | |
| amitk | Keybuk: rules.d/0-common-vars.mk:skipdbg = true | 14:03 |
| * ogra thinks he remembers a complaint from elmo about archive space | 14:03 | |
| ogra | (referring to linux debug stuff) | 14:04 |
| ogra | aha | 14:05 |
| ogra | * Remove debug meta packages since they are no longer publicly built (they | 14:05 |
| ogra | are now ddeb's) | 14:05 |
| ogra | -- Ben Collins <ben.collins@canonical.com> Sat, 26 Jul 2008 15:05:00 -0400 | 14:05 |
| ogra | linux-crashdump should pul in the necessary stuff as i understand the changelogs | 14:06 |
| rtg | ogra: I found an lpia device using AR2424, but I can't run 2.627 on it 'cause I don't have support for the display. | 14:06 |
| laga_ | ssh? | 14:07 |
| ogra | rtg, psb chipset ? | 14:07 |
| rtg | ogra: yep | 14:07 |
| ogra | vesa should work, just make sure you uninstall the xserver-xorg-video-psb package | 14:08 |
| ogra | it was a dep of xserver-xorg-video-all until recently and braks the display | 14:08 |
| ogra | *breaks | 14:08 |
| Keybuk | ogra: ah, well caught | 14:32 |
| Keybuk | interesting case that though | 14:33 |
| Keybuk | since the ddeb is a dependency of something in the archive ;) | 14:33 |
| Keybuk | quest scott# stap -e 'probe begin { print("hello world\n") exit() }' | 15:02 |
| Keybuk | hello world | 15:02 |
| Keybuk | \o/ | 15:02 |
| pitti | hi | 15:40 |
| pitti | Keybuk: you called? (sorry, I'm late, went to the supermarket) | 15:40 |
| Keybuk | pitti: interesting thing | 15:41 |
| Keybuk | systemtap depends on linux-image-debug | 15:42 |
| Keybuk | which is no longer in the archive, but a ddeb now | 15:42 |
| pitti | oh, is that the reason why it doesn't want to work out of the box? | 15:44 |
| pitti | (although the problem seemed to be much more general) | 15:44 |
| pitti | I can't even get the "hello world" example to run | 15:44 |
| Keybuk | you need an updated systemtap from Debian and linux-image-debug | 15:45 |
| Keybuk | then it works | 15:45 |
| pitti | ah, nice to know | 15:47 |
| pitti | but our -dbgsym doesn't, I take it? | 15:47 |
| Keybuk | we don't have a -dbgsym? | 15:47 |
| Keybuk | I was more asking why the deb is now a ddeb and on the other archive | 15:48 |
| pitti | does our pkg-create-dbgsym and the normal objcopy even work with the kernel? | 15:48 |
| pitti | Keybuk: well, I remember that there was a discussion about this once, but so far p-c-d doesn't have any special knowledge about kernel debug symbols | 15:48 |
| pitti | I wasn't really aware that it got dropped | 15:48 |
| Keybuk | it hasn't been dropped | 15:49 |
| Keybuk | err, let me back up | 15:49 |
| Keybuk | there's a package, called linux-image-debug-$flavour-$version | 15:49 |
| Keybuk | this contains the debug symbols for the kernel and its modules | 15:50 |
| Keybuk | it _used_ to be in the archive | 15:50 |
| Keybuk | but now, it's on ddebs.ubuntu.com (and named *.ddeb) | 15:50 |
| Keybuk | http://ddebs.ubuntu.com/pool/main/l/linux/ | 15:50 |
| pitti | ooh, I think I remember | 15:53 |
| pitti | Keybuk: the discussion, AFAIR, was that this is very big, and they were concerned about mirror size | 15:54 |
| Keybuk | right | 15:54 |
| pitti | so the linux source package was changed to name it .ddeb instead of .deb | 15:54 |
| Keybuk | so this is a general dependency of systemtap, in that you can't use it without it | 15:54 |
| pitti | so that our buildd -> ddeb.u.c. magic would pick it up and place it there | 15:54 |
| Keybuk | which also means that we can't have anything in our archive that build-deps on systemtap :-/ | 15:54 |
| pitti | it could only recommend -debug? | 15:55 |
| pitti | after all, people who use systemtap often build their own kernels anyway | 15:55 |
| Keybuk | well, it can't depend or recommend it at all | 15:56 |
| Keybuk | but it's useless without it | 15:56 |
| pitti | so, if the kernel team wants to keep the -debug.ddeb, it probably needs to be replaced by some runtime warning then? | 15:57 |
| pitti | "Enable ddeb archive and install -debug, kthxbye" | 15:57 |
| pitti | I'm not entirely sure about why -debug was moved to ddeb., just *how* it was done | 15:57 |
| Keybuk | it would be kinda cute to replace bootchart with a systemtap script | 15:58 |
| Keybuk | then we'd get really high quality reporting, with analysis of system calls, etc. | 15:58 |
| pitti | would be pretty expensive, though, if it really requires -debug | 16:01 |
| pitti | BenC, rtg: was there any other reason to move -debug to http://ddebs.u.c. than mirror space saving? | 16:03 |
| rtg | pitti: dunno, I wasn't involved in that one. | 16:04 |
| BenC | pitti: basically consistency with out debug type things | 16:04 |
| pitti | ok | 16:05 |
| pitti | If this runtime "proper error message" thing is feasible, that would certainly be helpful | 16:05 |
| pitti | also for people who build their own kernels, to tell them what they need to change to make systemtap work | 16:05 |
| rtg | I think its faster to reformat a drive then it is to remove a 40GB ccache directory | 16:09 |
| laga_ | 40GB ccache sounds scary | 16:10 |
| rtg | laga_: i386/amd64 caches for every release since dapper | 16:10 |
| Keybuk | pitti: only expensive to compile | 16:11 |
| Keybuk | once built, it's just a kernel module | 16:11 |
| pitti | Keybuk: I meant expensive to download | 16:16 |
| Keybuk | pitti: ? | 16:16 |
| pitti | Keybuk: if you need to download and install some 200 MB of -debug image to use systemtap, that's expensive | 16:16 |
| Keybuk | sure | 16:16 |
| Keybuk | but that's what you need to download :) | 16:17 |
| Keybuk | well, that's what you need to download to develop and compile systemtap modules | 16:17 |
| pitti | right | 16:17 |
| jdong | is there another kernel upload scheduled before the final release? | 17:12 |
| jdong | if so, I'd like to beg for the pulling of some ath9k patches from wireless-testing.git | 17:13 |
| jdong | particularly, the ANI enable patch and oops due to spinlock typo | 17:14 |
| rtg | jdong: such as? | 17:22 |
| jdong | rtg: commits 7768a47e04750e128ae97bbc759a2f5bc60ba668 and a6ebef8b1933be9eb867858ca11dee3c87da896c in particular | 17:28 |
| jdong | rtg: the spinlock typo causes kernel oopses relating to ath9k, and without ANI ath9k gets only about half the effective range as madwifi in my testing | 17:28 |
| rtg | jdong: luis didn't get those in under the regression policy? Seems like an oops would qualify | 17:29 |
| jdong | rtg: I'm not sure about the timeline; I don't see it in linux-2.6.git yet | 17:31 |
| rtg | jdong: I'll talk to linville about the first one. | 17:31 |
| rtg | jdong: I'm going to almost certainly have a compat-wireless backport by the time Intrepid releases. | 17:32 |
| jdong | rtg: sounds good; thanks very much for your time :) | 17:32 |
| rtg | jdong: is there a bugzilla report on the oops? | 17:42 |
| jdong | rtg: not sure, but here's the ath9k-devel thread on it: https://lists.ath9k.org/pipermail/ath9k-devel/2008-October/000407.html | 17:44 |
| === BenC1 is now known as BenC | ||
| jdong | linville was involved | 17:44 |
| rtg | jdong: well, there was no confirmation at the end of the thread, so he probably forgot about it. | 17:46 |
| rtg | I just talked to him, I think he would get it in as a regression if it had a bugzilla. care to create one? | 17:47 |
| jdong | rtg: well I've never filed a kernel bug before; I'm not sure I can get the format right the first time. | 17:49 |
| mathiaz | Hi - why is that linux-image-2.6.27-6-virtual ships ./boot/vmlinuz-2.6.27-6-server (and modules in ./lib/modules/2.6.27-6-server/) ? | 18:15 |
| erichj | is bug #260675 going to be addressed in time for 8.10 release? | 19:40 |
| erichj | https://bugs.launchpad.net/ubuntu/intrepid/+source/wacom-tools/+bug/260675 | 19:40 |
| rtg | erichj: it'll make the release. | 19:47 |
| erichj | rtg, thanks | 19:47 |
| === chuck__ is now known as zul | ||
| foka | Hi! I recently noticed that the Ubuntu kernel is compiled with CONFIG_X86_E_POWERSAVER turned off, | 21:57 |
| foka | Thus Ubuntu installed on VIA C7 notebooks do not enjoy cpufreq powersaving. | 21:57 |
| foka | Is that on-purpose, or an oversight? (Debian kernel has it compiled as a module). Thanks! | 22:00 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!