[03:16] i'm having a hell of a time finding the source tree for the EXACT kernel I'm running [03:16] i'm running 2.6.27-9-server, and every package I try to get installs source for 2.6.27, 2.6.27.2, or some variant thereof [03:16] none of them are -9 [03:17] I'm running 8.10-server [03:17] can anyone help me? [03:18] i've tried apt-get install linux-source, apt-get source linux-image-2.6.27-9-server, and I even tried the .deb package directly i found on ubuntu's software repository that claimed -9 [03:18] none of which compile as -9 though [03:26] x86: apt-get source linux [03:27] x86: If you just compile with make, you won't get -9. That's passed as EXTRAVERSION in the package build scripts. [03:27] mjg59: well I need -9 appended and .2 taken off [03:27] x86: Then use the package build scripts [03:28] mjg59: when I compile my Digium hardware drivers, it puts them into /lib/modules/2.6.27.2, instead of /lib/modules/2.6.27-9-server like everything else [03:28] mjg59: got a tutorial? [03:28] Wait. Why are you building against the kernel source rather than the kernel headers? [03:28] Don't do that [03:28] it requires the kernel source tree [03:28] (it needs /usr/src/linux) [03:28] Then it's broken [03:29] it uses genksyms and such [03:29] Unless it's actually sourcing C files into the build, all it needs is the linux-headers-2.6.whatever-foo package [03:29] stuff in /usr/src/linux/scripts [03:29] Which includes everything needed to build modules [03:29] Including genksyms [03:31] ok, so I can ln -sf /usr/src/linux-headers-2.6.27-9-server /usr/src/linux ? [03:32] hmm, looks like! :) [03:33] Well, or make the drivers build against /lib/modules/`uname -r`/build rather than /usr/src/linux, but yeah [03:34] excellent [03:34] you're a lifesaver thanks so much :) [03:36] No problem === lamont` is now known as lamont [12:06] amitk: do you have the need for madwifi on armel? === BenC1 is now known as BenC [12:39] hi [12:39] does stuff like [12:40] [ 770.458139] end_request: I/O error, dev sr0, sector 16679280 [12:40] [ 770.458158] Buffer I/O error on device sr0, logical block 2084910 [12:40] mean my dvd drive is screwed? [12:40] i get it every time i insert a disc [12:40] but e.g. a dvd movie seems to play fine in mplayer [12:40] it seems to be only when i insert a disc [12:41] it happens with 2 drives i've tried [12:50] BenC: no [14:12] BenC: what do you think about moving drbd to dkms? [14:12] zul: for what purpose? [14:13] so the kernel and userspace matches up if there is a need to change in the release cycle [14:13] zul: what is drbd? [14:13] zul: we do a lot of uploads through the development cycle...not sure how much more control you would need [14:14] zul: separating it would require you to add some dependency to make sure upgrades work [14:14] rtg: its cluster program which allows you to share disks [14:15] true...which version did you port to 2.6.28? [14:15] zul: the same as in intrepid....we can pull whatever updates you want though [14:15] BenC: ok thanks.. [14:16] that makes life easier [14:24] apw: I see that Jaunty has EDAC_AMD76 enabled. Does it have the same issue with AGP apertures as Intrepid? [14:34] BenC: are you working on Jaunty linux-meta? I got pitti to flush the kernel, so LRM ought to complete sometime today. [14:36] rtg: Yeah [14:36] rtg: thought you were leaving today? [14:36] BenC: in awhile. [14:36] gotta produce a blueprint first [14:38] BenC: pgraner is bored and listless in an airport with his CPU running full on. Maybe you could look over his shoulder and figure out why his update to -proposed isn't working. [14:57] rtg, i am unsure if .28 is affected by the agp thing, i wonder if we have any h/w which might demonstrate that [14:58] it seems like a very northbridge specific problem. [14:59] or is it southbridge, I forget. [15:00] the reporter says nothrbridge specific [15:02] apw: how about bugging the maintainer to see if it sounds like a familiar problem? [15:02] yeah ... that is a good suggestion [15:34] hello, I have a Q.; I'm maintaining custom kernel with IMQ patch for QoS, and using the guide at wiki.ubuntu.com for building custom kernel, but can't find where to change revision numbers for builded kernel, I've looked at debian/rules but can't find it.. [15:42] lod__, the version number comes from the debian/changelog [15:43] from the first line or? [15:43] lod__: use dch or dch -i to manipulate it [15:45] it opens changelog for editing? what should I change? [15:47] i think i've got it [15:47] 10x [15:48] it might take 1-2 iterations to get it fully right, there are certain numer/name combinations the kernel buildscript won't like [15:49] if I inc. the last number, and made no change to ubuntu-modules, will It compile? [15:52] Is there a reason to add jaunty? linux (2.6.24-22.45ubuntu1.1) jaunty; urgency=low [15:55] it will compile, but you'll get problems [15:56] when there's a new ubuntu kernel it will overwrite yours [15:56] I suggest naming it differently [15:57] when there is new kernel I'll recompile it with the patch [15:58] depends how big your audience is.. if it's only for you it's surely ok whatever you do [15:58] I just don't wan't to override mine even If there's no change, like now.. [15:59] It's for personal use.. Not that I'll forbid someone to use It.. [15:59] 10x again.. [16:00] have a great day.. [16:47] rtg you acked one of my threads on the email list (the CONFIG_EDAC_AMD76X bug), are you acking the patch option or the blacklisting option? [16:48] apw: um, I thought I only saw one, i.e., disabling the config option. [16:48] in the leader i also suggested we can blacklist the module, though i know not where those come from! [16:49] (blacklisting as an alternative) [16:49] I wonder how many users are afflicted with this problem? [16:50] not a huge number i'd guess from the bug [16:51] i am guessing we have a fair number of AMD based systems in our team alone and noone there had the combination of h/w which triggered it [16:51] apw: you've suggested blacklisting to them? [16:51] they found blacklisting worked yes [16:51] (the submitter) [16:51] what is the downside for folks that have this northbridge, but don't have the radeon GPU? [16:52] the downside of blacklisting by default, I meant [16:52] well its not just radeon, the apature is not setup correctly so no AGP graphics will work in DRI mode [16:53] the edac module for them is supposed to report on correctable errors that occur on the system [16:53] and they obviously didn't have it in hardy [16:53] I assume you have to have ECC RAM before EDAC makes sense? [16:54] that is my expectation yes, i don't know a lot of the details sadly [16:54] so, that likely means the effects are primarily in the server community. [16:55] yes that was my take [16:55] i wonder if it would [16:55] be better to leave it built, but mark it blacklisted by default [16:55] I think in that case that blacklisting makes the most sense. It leaves the module available for installation if its really needed, and afftects the smallest number of platforms. [16:56] ok, can do that ... where are our blacklists ? i assume they are in some other package? [16:56] modules-init-tools [16:56] actually lets restate that ... do we default it blacklisted or just leave it as is [16:57] you add a blacklist file in module-init-tools, leave the config option enabled. [16:57] ok ... so i'll look at that then. a good excuse to figure out how one updates m-i-t [16:58] how does one even find the owner of a package [16:59] Ubuntu owns the local modifications to the package. this isn't something upstream is likely gonna want. [16:59] apw, A quick question. Just saw your mail about the audio patch. I want to make sure I understand: you booted an unpatched hardy kernel and that worked ? [17:00] smb_tp, yes that is correct. i booted the latest .24 that was installed before i updated and that the sound is routed correctly. it appears the specific system is not recognised back there, and the default is much closer to useful than that in intrepid [17:00] Interesting. [17:01] smb_tp, apw: its not surprising. I think there were a number of ALSA regressions between Hardy and Intrepid. [17:01] Ok, but this way. BTW, the STAC_DELL_M6 is defined there... Not sure this matters [17:03] apw@dm:~/git2/ubuntu-hardy$ git grep DELL_M6 [17:03] apw@dm:~/git2/ubuntu-hardy$ [17:03] i don't see it in my tree? [17:03] lum? [17:03] ahh point [17:03] yeah, hardy LUM [17:04] possibly it defaults there to something sensible and less sensible in intrepid. [17:05] so when i updated to intrepid would hardy-lum have been deinstalled [17:05] no, all is in the modules dir for the kernel. the co-exist [17:05] s/the/they/ [17:05] ie. i no longer have a full modules install for my hardy kernel [17:06] apw: yep, it likely did an autoremove when you dist-upgraded - furthermore, we're using ALSA straight from the kernel in Intrepid. [17:06] oh yeah there they are, ok so this testing is with lum installed, so i guess its safe to say that the default was better in hardy, so any backport there seems unessary risk [17:06] right [17:07] ok so i think i've done the right thing and done nothing there, and marked the nomination invalid with a reason it works there [17:08] rtg is it correct to say that the kernel team owns module-init-tools? [17:08] Right, yes. I actually just wondered about the fact it worked without changes (and that the model would be missing). [17:09] smb_tp, they have definatly added support for the whole family of dells to the kernel, so i am not supprised that this one got broken [17:09] they added a general 'reference' implementation for ICH9 [17:10] which was not there before, and the refereence is wired dfferent to the dell i have [17:10] so it needs a model quirk to fix that ... [17:10] Ah, ok. Well it seems everything is wired differently nearly everywhere [17:11] its mad as mr mad jack mc mad, basically you need 20 binary words for every machine in existance in the kernel [17:11] soooo scalable [17:29] rtg so i have gotten the module-init-tools source package and modified it to add the blacklist entry -- debdiff seems to show something sensible [17:29] what does one do next? [17:30] i guess i should upload them to my PPA and see what happens, if they build ok [17:31] apw, right, and if you are happe you need to find someone with upload rights. [17:31] s/happe/happy/ [17:31] * smb_tp curses his fingers [17:32] smb_tp, thanks ... like a sponsor i guess [17:33] Well in this case rtg might be a good choice. ;-) [17:35] And after that it goes the usual way through archive admins. This package might be special since so closely related to kernel. In the past I tried to look at the changelog and find out who did previous uploads for Ubuntu [17:40] smb_tp, yeah rtg and ben are the people who did it before so i'll go to them [17:41] _once_ i have figured out how to put it in my PPA [17:41] smb_tp, got any how to's for PPA's ? [17:42] smb_tp: [ 52.835985] usb 1-1: airprime converter now attached to ttyUSB0 [17:43] should be quite simple. change changelog to something with an extension to recognize. release should be something without -proposed or so. then there was a wiki how to configure dput or dupload. [17:43] ^^ the sparc thingy.. I'll prepare a tree for you tomorrow or so [17:43] smb_tp: based on what .git tree do you want that? (hardy) [17:43] struggling to find any docs at all, which is unlike us [17:44] fabbione, Ok, sure. I remember. Was that missing in the hardy tree? Oh well, I figure out somehow. :) [17:44] apw, Jus a sec. I try to find it... [17:45] smb_tp: when you enable the usbserial, it asks tons of config options. there are 3 needs to be set. I'd prefer to send you a diff or something but I need to know on top of git tree you want it [17:45] smb_tp: -updates? -security? whatever? [17:45] fabbione, -master is fine [17:45] smb_tp: ok [17:46] if then you want to propagate the change forward, it's up to you. I only care about hardy for this machine :) [17:46] because I don't plan to reinstall it anytime soon [17:47] Sure. :) NTARS [17:47] NTARS? [17:47] Never Touch A Running System [17:47] ah ok [17:47] yeah indeed [17:48] smb_tp, found it ... finally [17:48] apw, Ah, beat me. I also could just send you my config, if you like [17:48] smb_tp: this same machine was the "one" that did bootstrap all of ubuntu sparc alone.. a single cpu (old as hell) with 512Mb of RAM.. it's undestroiable [17:49] * fabbione can still hear elmo's screaming :) [17:49] LOL, I can imagine. How long did that take? [17:50] it took about 3 weeks to catch up with the archive [17:50] and then it was just keeping up [17:50] nasty [17:50] it did miss hoary release for one package because of my mistake. [17:50] The machine or the PPA upload? [17:50] smb_tp, please do send my your archive if you don't mind [17:50] the machine, is nasty :) [17:51] apw: it's a very small Netra T105.. google for it :) [17:51] it was doing OOo builds over nds [17:51] nfs [17:52] it doesn't have enough disks for that :) [17:52] * apw feels for the machine facing that life [17:53] apw, sent you the config [17:53] which is probably most that is special [17:54] You either use dupload (simpler since this is a config for that) or dput [17:55] Since you used debdiff, you already have a source package for it [17:55] i seem to have dupload installed and your config is for that, so ... seems ok [17:56] yep learned that side of things when trying to follow along with the linux-meta update [17:57] just change the name ;-) [18:00] smb_tp, yep .. thanks ... i've hit upload and it did something reasonable looking -- how long does it take from there to appear in your ppa? [18:01] apw, You will see it quite fast. Then it takes some time to build (well the kernel at least) mit should be better [18:01] ahh there it is .. cool [18:01] last i looked there were no PPA builds on the buildd's, of course _now_ they are all rammed. typicial! [18:02] dum di dum... [18:02] one could be me [18:06] yeah one is you! [18:06] * smb_tp sneaks away [18:34] rtg ok ... i have and updated module-init-tools which seems to work. i have built and uploaded it to my PPA and installed it from there locally and good things happened. you available to talk me through the next step? [18:48] smb_tp: works fine .. all of it [18:48] smb_tp: i am on ppp over cdma over usbserial on sparc... wanna check? :P [19:15] fabbione, that sounds like some kind of mental torture [19:16] apw: i never claimed NOT to be in some kind of BDSM games with my computers [19:17] heheh, and who is the masters and who the slave? [19:17] i rule the world and you? ;) [19:17] i like to think so, but suspect i am also deluded === doko__ is now known as doko [19:23] ehhee [20:24] apw: I'll catch up to you tomorrow wrt to module-init-tools.