[01:24] hey all [01:25] im thinking of grabbing 2.6.25 kernel for built in 8165 wlan support that i hear rumors of [01:25] is there an official 2.6.25 (k)ubuntu kernel package? [01:27] no. Do you mean 2.6.27? [01:28] i didnt mean that [01:28] crimsun: what if i mean that now [01:28] ? [01:30] (the only 2.6.25ish one is for -ports) [01:41] crimsun: i didn't explain myself well. i dont care what version the knerel is, so long as it has a module for the 8165 wlan chipset [01:41] but i found some ubuntu docs [09:49] cking: I had a git tree from earlier in Intrepid, and just did a 'git pull' [09:49] this gave me a bunch of conflicts, though I never modified by working tree. why is this? [09:49] mdz: I suspect it's because it's been rebased. [09:53] I see this occur when ever the latest rc patches are pulled in - if there are too many of these merge conflicts I just get the entire repo again from clean - which is probably not an efficient way of working buy my network connection is fast [09:53] s/buy/but [09:53] I am sure amit has the necessary runes to resolve this efficiently [09:53] mdz: during development, git pull will almost never work [09:53] mdz: because we rebase, as cking mentioned [09:54] amitk: how do you keep your tree up to date? [09:54] git fetch; git rebase origin [09:54] (I won't ask why we rebase...) [09:54] but wait [09:54] since your tree is probably in a mess right now... [09:55] mdz: do you have any local changes you care about? or do you just need the latest kernel tree? [09:57] mdz: 'git fetch; git reset --hard origin' will restore you to a pristine tree if you don't have (or care about) local changes. [09:59] amitk: thanks [10:12] Hi I'm looking at bug 280451 raised by doko - is there any historic reason why the ports kernels builddep on gcc-4.1 specifically? === asac_ is now known as asac === thegodfather is now known as fabbione === thegodfather is now known as fabbione [14:35] Hi I'm looking at bug 280451 raised by d o k o - is there any historic reason why the ports kernels builddep on gcc-4.1 specifically? [14:36] I just have this recollection that I saw a check in /Makefile ages ago that enforced use of gcc 4.1 for a certain arch but I can't see it [14:36] there now [14:36] munckfish: some of the arches used to have compiler dependencies. dunno if they are still true. [14:37] compiler version dependencies, that is. [14:37] rtg: yep [14:37] if they no longer exist in Hardy, then they are probably not appropriate for the ports kernels. [14:37] https://bugs.launchpad.net/ubuntu/+source/klibc/+bug/280451 [14:38] munckfish: propose a patch and get smb_tp to test build [14:38] rtg: yes will do [14:42] munckfish, best mail to the kernel team list [14:43] munckfish, if you are done and have something to pull [14:43] smb_tp: yep usual procedure. [15:39] is there some news on the e1000 for module regarding the eeprom-corruption? [15:42] njpatel: the EEPROM corruption is related to ICH{8,9,10} e1000e parts, and yes, it appears to have been resolved by marking mapped memory to be read-only. === BenC1 is now known as BenC [15:44] rtg: anyway I can do that on my laptop right now? I am at the GNOME/GUI-hackfest without any network-access on my own laptop (just using a colleagues machine for the moment) [15:47] njpatel: you can do what on your laptop? update? 2.6.27-5 and later should be fine. [15:48] rtg: oh... hm... I already updated to 2.6.27-6 (via usb-stick) and that still does not give me back a working eth0 [15:49] rtg: i am on a core 2 duo (using amd64 arch) [15:49] njpatel: what NIC? lspci -vvnn [15:49] rtg: using a Ubuntu 7.10 live cd ethernet works so the hardware is ok [15:51] 00:19.0 Ethernet controller [0200]: Intel Corporation 82566MM Gigabit Network Connection [8086:1049] [15:52] Subsystem> Lenovo Device [17aa:20de] [15:52] njpatel: hang on a sec... [15:54] njpatel: ok, that is an ICH8 device. do you see any eth devices with 'ifconfig -a' [15:55] rtg: I get lo and pan0 [15:55] njpatel: look in dmesg, e.g., dmesg | grep -i e1000e [15:56] oh one sec [15:56] njpatel: also, make sure its not still black listed in /etc/modprobe.d [16:00] rtg: nothing shown from dmesg [16:00] njpatel: check for black listing? [16:01] rtg: there is blacklist-e1000e in /etc/modprobe.d still [16:01] just delete that? and try to restart networking? [16:02] njpatel: ok, then you're not completely up to date. 'sudo rm /etc/modprobe.d/ blacklist-e1000e; sudo update-initramfs -u' [16:03] probably due to me just updating the package manually and not via apt-get or synaptic (since I have not network) so some scripts were skipped [16:04] rtg: manually loading e1000e now should work I assume [16:05] njpatel: well, my first attempt at removing blacklist-e1000e in module-init-tools wasn't correct. Steve Langasek had to come along behind and clean up my mess. [16:05] njpatel: yep - you should be able to just modprobe e1000e (or reboot) [16:06] hm.. looks good sofar... trying to plug in a cable an see NetworkManager hopefully do its magic [16:07] rtg: sweet works again [16:07] rtg: you don't also happen to know a bit about iwl3945 [16:07] :) [16:08] njpatel: a bit, though I'm not sure I'll of much help. [16:08] s/of/be of/ [16:14] rtg: ok huge thanks at least ethernet is working again [18:36] hum [18:36] aufs is broken *again* [18:41] laga_: in what way? [18:41] NFS branches [18:41] i'd like to blame kees cook commit (118463fa8236a123fdf6f81fb0c0176f7fd1caa7), but i can't tell for sure yet [18:41] i still wonder why i have to constantly explain funny changes to aufs to junjiro [18:42] kees cook re-introduced aufs_do_setattr which hasn't existed in a long time in aufs. although it looks like it's just a wrapper, so i'm not sure if it's to blame [18:43] amitk: http://launchpadlibrarian.net/18342118/DSC00183.JPG - this is what you get if you try to boot a mythbuntu-diskless client [18:46] i will set up a test system tomorrow and revert that commit. i'm glad someone noticed before the release :) [19:02] laga_: just talked to kees. If reverting that patch fixes the problem, could you possibly prepare a patch separating the vfs changes from the fsetattr changes? [19:03] yes, i think i could do that. i'll get back to you [19:03] do we have any way to support RT2700E wlan chipsets ? [19:03] apparently they are in the eee1000 [19:18] ok [19:19] following -> Intrepid crashes on my Intel D945GCLF + Atom 230 (its a Intel bundled mobo+processor) [19:19] desktop / server (also tried alternative version) [19:22] you should mention that you try to run 64bit kernels on it, i doubt anyone will actually associate 64bit with atom (since thats the exact thing you ran into in the other channels ;) ) [19:22] im currently installing the 32bit version of Intrepid beta server (32bit), the installation is nearly completed [19:22] ogra: i wasnt done typing yet :P [19:22] heh [19:24] So -> problem is, Intrepid (64) kernels do not work on the Atom 230 (probably not on the newer 330 either, i'll buy this as soon as possible) [19:24] but i cannot test that yet [19:24] funny detail, with acpi=off .. the kernel of the livecd's does work [19:27] What can i do, how can i help with adding support / fixing what is broken? [19:28] amitk, did you ever do any work with 64bit atoms yet ? [19:29] ogra: not yet... so what does 'crashed' mean here? [19:31] kernel segfaults, no CPU time for xxSeconds [19:31] errors [19:33] mrxmike: It will be hard to debug or look into without detailed info. dmesg, lspci -vvnn, /proc/version_signature attached to a bug report would help a lot. [19:35] but booting with acpi=off seems to suggest acpi bug in bios or kernel [19:44] amitk: i will provide all those details as good as i can, its easy ... as i can use the livecd for it... [19:45] but i just installed 32bit server, and i do want to use my server as well .. (so i keep that on) [19:46] mrxmike: so it works just fine with 32-bit? [19:46] i think it does yes [19:46] (well quiet sure actually) [22:22] Hi, on intrepid, why the current version of linux is not the 2.6.27 on powerpc ? [22:32] lafeuil: because powerpc is now community supported and therefore uses the linux-ports kernel source not the standard one [22:33] it's stuck 2.6.25 cause no one was motivated enough or had enough time to upgrade it [22:33] I'm hoping that'll change in intrepid+1 [22:33] cause I'm hoping to do something about it myself [22:35] munckfish: Ok thanks for your answers [22:35] lafeuil: np [22:35] When is intrepid+1 ? [22:36] well pretty much 6.5 months from now I guess [22:36] lafeuil: is there a specific feature you're missing? [22:37] munckfish: no, I understand now [22:38] munckfish: thank you [22:38] bye [23:07] hi, did anybody test gspca with 2.6.27? in worst case it does not even boot with webcam connected... [23:09] 2 ppl dont get the webcam working with it... with 2.6.26 old gspca it works [23:14] hey all... im reading the ubuntu custom kernel docs... i dont want to use the git method on a production kernel, do i? i basically want a kernel built the same way as the default kernel, + the 8159 wlan driver... [23:15] just compile it externally,what the problem [23:17] Kano: isnt there a certain kernel package out there just for (k)ubuntu? [23:17] you only need the headers + a driver that compiles against the kernel [23:17] with whatever relevant patches that we use [23:17] Kano: the driver is in the kernel id like to use [23:18] so i just grab any linux-headers and kernel, and i dont worry about patches or anything like that? [23:18] which config option? [23:18] Kano: its the wlan module for 8159 chipsets [23:18] config option? [23:19] there is no option with 8159 in the name.. [23:19] Kano: i dont know, i head someone say it was in there so i was gonna look through menuconfig [23:19] i think you are dreaming [23:19] check 8150 [23:19] and pinch me [23:19] CONFIG_USB_RTL8150=m [23:19] is already set [23:19] does it add support for 8159 in the newest kernel or something? [23:20] /sbin/modinfo rtl8150 [23:20] shows it, what i find problematic is that rt2500usb and rt73usb share the same ids [23:22] that doesnt look like it supports the 8159 chipset anywya [23:22] lsusb [23:23] then you can check the id [23:24] its not a usb nic [23:24] then use lspci -nn [23:24] and show the line [23:24] Kano: im beginning to think i was misinformed.. sure, one sec [23:24] 8150 IS usb [23:25] also wlan is very simple to use with ndiswrapper [23:25] im a knuckle head, its 8185 [23:26] and i have serious problems with ndiswrapper, which im certain constitute a bug [23:26] http://ubuntuforums.org/showthread.php?p=5931608 [23:26] which is why im here [23:27] administrator_: rtl8180 is already there, if you want to use ndiswrapper you have to blacklist it [23:28] i dont have ndiswrapper, but i didnt have wlan0, so i assumed i needed it [23:29] administrator_: grep 10ec /lib/modules/$(uname -r)/modules.pcimap |grep 8185 [23:29] sorry, i meant to say 'i dont want to use ndiswrapper, im only using it because i didnt have an interface for my wlan, so google searching seemed to imply i needed it_ [23:29] ndiswrapper is usually very good [23:29] but in that case you would have to blacklist rtl8180 [23:30] Kano: nothing returns from that grep [23:30] then your kernel is too old [23:30] i have 8139 and 8169, which are obviously wired drivers [23:30] i use 2.6.27-6 [23:31] Kano: and you have it right? [23:31] of course [23:31] great, sorry for the confusion, my fault [23:32] so i can just grab linux kernel sources from anywhere right? there isnt any 2.6.27 *buntu specific-patched kernel that you would recommend i use [23:32] in 8.10? [23:33] it is [23:33] these need to go into production, asap, or i would just use intrepid personally [23:33] i need a kernel for use with 8.04.1 [23:51] Kano: thanks, those .debs worked great [23:51] fine [23:51] lol