[12:05] dilinger: btw, would you happen to know: debian/2.6.12/ia64: working or no? === dmk [n=dmk@host81-154-88-175.range81-154.btcentralplus.com] has joined #ubuntu-kernel === ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | http://www.ubuntulinux.org/wiki/KernelTeam | The kernel is not in, please leave a message after the beep...OOPS, unable to handle kernel paging request | http://people.u.c/~lamont/Archives/kernel-team@ubuntu.com--2005/ playground: kernel-debian--preX,15--2.6.12 === mkrufky [n=mk@68.160.103.77] has left #ubuntu-kernel [] [12:36] lamont? [12:37] si? === Seveas [n=seveas@seveas.demon.nl] has joined #ubuntu-kernel === Seveas [n=seveas@seveas.demon.nl] has joined #ubuntu-kernel [12:41] nevermind :) [12:42] lamont: do you know anyone besides neuro and elmo that handle the debian buildd's? [12:43] depends on what you mean by 'handle' [12:43] I do hppa/ia64 [12:43] someone that can setup auric as a sparc buildd [12:43] infinity is involved in m68k [12:43] sparc is elmo, iirc [12:43] ok [12:43] and auric would certainly be DSA-types [12:44] elmo has been MIA all day === Seveas [n=seveas@seveas.demon.nl] has joined #ubuntu-kernel === Seveas [n=seveas@seveas.demon.nl] has joined #ubuntu-kernel === dmk [n=dmk@host81-154-88-175.range81-154.btcentralplus.com] has left #ubuntu-kernel ["Leaving"] === dmk [n=dmk@host81-154-88-175.range81-154.btcentralplus.com] has joined #ubuntu-kernel === dmk [n=dmk@host81-154-88-175.range81-154.btcentralplus.com] has left #ubuntu-kernel ["Leaving"] === mkrufky [n=mkrufky@user-12lcl1s.cable.mindspring.com] has joined #ubuntu-kernel === Bicchi [n=chatzill@adsl-065-006-167-199.sip.mia.bellsouth.net] has joined #ubuntu-kernel === Bicchi [n=chatzill@adsl-065-006-167-199.sip.mia.bellsouth.net] has left #ubuntu-kernel [] === zul [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [04:10] evening [04:10] yo [04:10] whats up [04:18] i think we could try updating spca5x driver this weekend ;) [04:49] BenC : We have 1001 offers for sparc buildds, several of which are much beefier than auric, but I need to talk to elmo so see if we can get any going. [06:38] morning guys [06:38] morning [06:38] BenC: sparc kernel is up :) [06:59] infinity: we have 2 buildd's now [07:00] infinity: and none of the offers I saw were worth anything [07:00] infinity: elmo already has auric going [07:00] fabbione: cool [07:00] auric is steadily building packages as we speak [07:12] BenC: when do you plan to put your e3k back online? [07:12] after I move [07:12] hopefully in the next week [07:12] ah cool [07:12] i was expecting a 2 months movement ;) [07:12] hmm i need to add the other buildd online [07:12] we got 2 offered, but one (pretty fast) keep crashing [07:13] i need to check on this on [07:13] +e [07:13] Setting up linux-image-2.6.12-9-sparc64 (2.6.12-9.14) ... [07:17] vultus5:/# debootstrap breezy /mnt http://mirror.int.fabbione.net/ubuntu === fabbione tests an installation === Kamion [n=cjwatson@83-216-156-196.colinw664.adsl.metronet.co.uk] has joined #ubuntu-kernel [09:11] BenC : Well, one offer I saw was for a machine pretty much identical to auric, but with 2G of RAM. [09:11] Anyhow, it's the matter's well in hand, I don't much care. :) === janimo [n=jani@Home03207.cluj.astral.ro] has joined #ubuntu-kernel === Kamion [n=cjwatson@83-216-156-196.colinw664.adsl.metronet.co.uk] has left #ubuntu-kernel [] === JaneW [n=JaneW@wbs-146-189-29.telkomadsl.co.za] has joined #ubuntu-kernel === lamont [n=lamont@dhcp-sn38-07.hrz.uni-oldenburg.de] has joined #ubuntu-kernel [11:30] Bestest kernel EVAR. === janimo [n=jani@Home03207.cluj.astral.ro] has joined #ubuntu-kernel [02:15] BenC: Uhm. Did you actually enable sk98lin in the .config? === Yagisan [n=jamie@60-240-38-8.tpgi.com.au] has joined #ubuntu-kernel [02:30] mjg59: that's a damn good question [03:00] and the answer would be no, but the upside is, 9.15 is going to upload in about 30 minutes [03:01] I forgot to enable ndiswrapper on amd64 aswell [03:01] Ha [03:01] BenC: Have you spoken to kamion? [03:02] no [03:07] BenC: Nothing should be uploaded to main at the moment unless you have clearance === zul [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [03:11] BenC: hold on.. wouldn't be an idea to do a test build around with that config change first? [03:11] are we sure that driver build everywhere? [03:11] sk98lin? [03:12] it will only be building everywhere skge and sky2 were building [03:12] well, that's probably everywhere except hppa, atm [03:12] that includes sparc too [03:12] I'm talking to Kamion first [03:12] hey [03:12] I know it builds on sparc [03:13] used it before on sparc64 [03:13] hey zul [03:13] BenC: hmm ok [03:13] should i run a test build for the sake of god? [03:13] hold a sec [03:13] sure [03:17] well, green light for 9.15 upload [03:17] fabbione: changes commited [03:17] BenC: did you add the drivers to the udebs? [03:17] so that also the installer can actually use it? [03:17] no, wasn't aware of that...let me do that [03:18] do you know how? [03:18] otherwise i can do it on the fly [03:18] got it [03:18] nic-modules, right? [03:19] is that the only place? [03:19] also, what does the question mark in "foo ?" mean in that file? [03:19] debian/d-i/shared/something... [03:19] foo <- the module MUST be there [03:19] ah [03:19] foo ? <- include it if it's there otherwise skip it.. no failure [03:19] so you must add a ? [03:20] since not all arches will ship this module, right? [03:20] right [03:21] ok, got that in too [03:22] updating the local tree [03:23] too about 15 minutes to collect the abi files for all 6 arches [03:23] took* [03:24] given that all of them do actually build. otherwise you are in trouble :) [03:24] oh crap [03:24] i forgot to switch archive [03:27] fabbione: so you said 2.6.14 should be done like 2.6.14-2.6.13.90-X.Y until 2.6.14 final, right? [03:27] no i suggest you use something like 2.6.13.90 [03:28] as major upstream [03:28] never put the final, until it's final [03:28] well, I was going to call it linux-source-2.6.14 [03:28] but the version would start at 2.6.13.90 [03:28] hmmm oh yeah [03:28] sorry.. that's right [03:28] ok [03:29] just be sure that other than source, there is nothing that shows up as 2.6.14 [03:29] actually going to do .92 for -rc2, .93 for -rc3, etc [03:29] right [03:29] otherwise people will harass you badly [03:30] BenC: it's a good idea after you branch, to do baz cachereb [03:30] meh [03:30] cacherev [03:30] otherwise it takes ages to switch/checkout [03:31] does changelog Closes entries actually close bugzilla bugs? [03:32] nope [03:32] bugzilla does nothing [03:33] ok, it's uploaded, I had to get it quickly so Kamion wasn't holding off Colony 5 forever [03:34] ah ok [03:34] well than i won't bother to build [03:34] we will see what the buildd will say [03:35] if it breaks, you better get ready to upload again :) [03:35] meh...i have to fix that audigy bug [03:35] for some reason, I knew I was going to have to do a 9.15 today [03:35] zul: what bug? [03:35] i have a audigy here... [03:36] BenC: it's always like this .. [03:36] specially when you approach release === fabbione finishes to prepare hoary-security [03:36] fabbione: some pci ids were removed from the audigy driver in 2.6.12 they are back in 2.6.13 [03:37] zul: ENOCARE :) works for m [03:37] +e [03:37] including remote control [03:37] and my boss uses audigy on his computer at home and he knows im on the kernel team [03:37] blah.. [03:38] zul: tsk :P [03:38] tell your boss to buy a better card [03:38] everyone that's on the "kernel team" needs to add themselves to the KernelTeam wiki page so I know who you all are :P [03:38] i did :) [03:38] BenC: he is on the kernel team [03:38] speaking of which [03:39] we need to clean it up a bit [03:39] I did a little [03:39] can you remind me the URL? [03:39] I don't see any names on the KernelTeam page other than fabbione and myself [03:39] wiki.u.c/KernelTeam [03:40] that's right... [03:40] soon to be only yoouuuUUUUU [03:40] but WTF [03:40] we had a completely different page [03:41] with a lot of links and shit [03:41] where did it all go? [03:41] there's a diff URL in the topic [03:41] http://www.ubuntulinux.org/wiki/KernelTeam [03:41] I didn't see that one [03:41] same page [03:41] that's the way I found it [03:42] only thing I changed was s/fabio/ben/ [03:42] yeah i know [03:42] it must have been when the wikis fetish BDSM guys did move to YET another WIKI [03:42] we had all the policy and crap [03:43] http://udu.wiki.ubuntu.com/LinuxKernelRoadmap [03:43] there's that too [03:43] yeah that's the specs from UDU [03:44] https://wiki.ubuntu.com/KernelTeamTasks?highlight=%28kernel%29 <- [03:44] just search for kernel [03:44] it will spit out shit [03:45] ok [03:45] need to get all those links in the KernelTeam page [03:45] KernelTeam is supposed to be the main page, right? [03:45] they were there :/ [03:45] yes [03:45] ok [03:46] if you can, try to kill the links on how to build custom kernels [03:46] hopefully with 9.15 out, and Colony 5 building, I can take some time to do that [03:46] ehehe [03:46] it can wait after Breezy [03:46] so we might be able to assimilate more people [03:46] I got a lot of stuff I need to document, so I don't lose/forget it [03:46] the whole ABI-changing-upload [03:47] do I still need to do klibc? [03:47] jbailey is looking at it [03:47] ok [03:47] klibc is not extremely important [03:47] because an ABI change doesn't break it [03:47] or purge it [03:47] d-i, linux-meta, and lrm should be done [03:47] it makes klibc unbuildable [03:47] yeah it's all done i think [03:48] btw, auric's been chugging away at packages [03:48] I think sparc is safe from the arch-police for etch [03:48] afaik sparc is going to be SCC because the installer didn't get any love in sarge [03:49] SCC? [03:49] the installer worked great for me from sarge [03:49] BenC: Secondary Crappy arCh [03:49] Eh? Sparc was working when I beat on it last year. [03:49] I did the auric install with a sarge tftp image [03:49] that's what i was told [03:49] so.. [03:49] relax :) [03:49] BenC: if we can get your machine to build breezy in a serious way [03:50] I think if sparc32 gets dropped, things will be a lot better [03:50] we can't keep feeling sorry for the sun4c folks [03:50] than i can dedicate my sparc for testing [03:50] Already done for breezy. =) [03:50] breezy doesn't support sparc32 [03:50] yeah, debian needs to do the same [03:50] BenC: you said you have like 6 CPU/6GB of RAM [03:50] netbsd runs so much better on the sparc32 hardware anyway [03:50] that could run but it self 5 instances of buildd :) [03:51] That was at the advice of the aurora and the slackware for sparc or wahtever they're called. [03:51] fabbione: yep, it can do some serious damage to a build list [03:52] anyone ever looked at taking advantage of smp on larger packages? [03:52] mozilla, openoffice, kernel, X, gnome, kde, etc. [03:52] It's hard to do predictable in a generic way. [03:52] gcc too [03:52] glibc does it. [03:53] kernel does [03:53] yeah, I saw that the kernel does [03:53] X can't [03:53] Even with autoconf? [03:53] and there is no need anymore anyway [03:53] It ought to be able to. [03:53] I know my e3k can -j a kernel build in like 5 minutes or less [03:53] jbailey: it's splitted in 3M pkgs.. [03:53] BenC: Is that with or without ccache [03:53] ? [03:53] BenC: eheheh [03:53] without [03:53] sweet [03:53] ehehe [03:53] with ccache is going to fly [03:54] and if you are cool with the setup, with shared ccache between chroots it flies even more :) [03:54] how hard is that to setup [03:54] trivial [03:54] install ccache [03:54] easy [03:55] lrwxrwxrwx 1 jbailey jbailey 15 2005-05-16 17:57 cc -> /usr/bin/ccache [03:55] lrwxrwxrwx 1 jbailey jbailey 15 2005-05-16 17:52 g++ -> /usr/bin/ccache [03:55] lrwxrwxrwx 1 jbailey jbailey 15 2005-05-16 17:52 g++-3.3 -> /usr/bin/ccache [03:55] lrwxrwxrwx 1 jbailey jbailey 15 2005-05-16 17:52 g++-3.4 -> /usr/bin/ccache [03:55] nice [03:55] lrwxrwxrwx 1 jbailey jbailey 15 2005-05-16 17:52 g++-4.0 -> /usr/bin/ccache [03:55] lrwxrwxrwx 1 jbailey jbailey 15 2005-05-16 17:52 gcc -> /usr/bin/ccache [03:55] BenC: when you are ready to do it, i will give you the chroots ready for you [03:55] lrwxrwxrwx 1 jbailey jbailey 15 2005-05-16 17:52 gcc-3.3 -> /usr/bin/ccache [03:55] lrwxrwxrwx 1 jbailey jbailey 15 2005-05-16 17:52 gcc-3.4 -> /usr/bin/ccache [03:55] BenC: with all the setup and stuff [03:55] lrwxrwxrwx 1 jbailey jbailey 15 2005-05-16 17:52 gcc-4.0 -> /usr/bin/ccache [03:55] Do that in ~/bin [03:55] Add it to your path. [03:55] BenC: you will only need to allocate space that can be bind mounted [03:55] jbailey: there are much easier way than that [03:56] if [ -d /usr/lib/ccache ] ; then [03:56] export PATH=/usr/lib/ccache:"${PATH}" [03:56] export CCACHE_DIR=/usr/src/.ccache [03:56] export CCACHE_NLEVELS=8 [03:57] if [ -n "$(which distcc)" ] ; then [03:57] export CCACHE_PREFIX="distcc" [03:57] export DISTCC_HOSTS="localhost foo bar" [03:57] fi [03:57] fi [03:57] no symlinks [03:57] no mess [03:57] and ready for distcc :) [03:59] debuild won't see that, I don't think will it? [03:59] it does [04:03] Yeah, /usr/lib/ccache is love. elmo pointed it out to me the other day, and I felt dumb for having created the links manually. [04:11] nice [04:25] great, the sk98lin driver is failing to build at all, has nothing to do with the arch [04:27] BenC: Hmm. My patch appeared to build here. [04:28] BenC: It's missing the -I statement pointing it at its includes [04:30] got it [04:30] it still had -Ia/drivers [04:30] no idea why [04:38] wow, that's broken [04:38] Do I smell ANOTHER upload? [04:39] fabbione: the sparc64 breezy tftp image told me that I didn't have a ethernet interface, but that I _did_ have a firewire device :) [04:39] which is odd considering that I have no PCI on this e3k :) [04:40] but I do have a happymeal card [04:40] BenC: no idea.. [04:40] i didn't test breezy installer at all [04:40] but all the modules should be there [04:40] yeah, I loaded it manually [04:41] BenC: for me the breezy installer is completely untested [04:41] so it might fry stuff [04:41] be aware [04:41] ok [04:41] if it works is a plus :) [04:43] is the smp kernel for sparc64 tested well? [04:43] we did test it only on one machine [04:43] the breezy installer must be using the broken version of libdetect [04:43] i don't have SMP here [04:43] it's not seeing my sbus devices [04:44] that bug was in debian too, but it got fixed [04:44] it depends when it got fixed [04:44] libdetect ignores > 1 sbus buses [04:44] was awhile ago [04:44] we should compare the versions [04:44] like 3-6 months [04:44] if the fix is not in and it's not intrusive, we might be able to push it [04:44] I'm having to manully load hme and esp modules [04:45] i see [04:45] annoying [04:45] it only affects machines like mine that have things on other than sbus0 [04:45] i guess that problem is only in the installer, right? [04:46] BenC: anyway i need to stop for one hour or so [04:46] my brain is melting [04:46] well, if it uses detect for initramfs, then it will be there too [04:46] no it doesn't [04:46] afaik at least [04:46] jbailey: ??? [04:46] i need to stop for a while [04:46] bbl [04:49] libdetect? [04:49] What's that? [04:50] Ubuntu's d-i uses hotplug / udev [04:50] what does it uses to scan the sbus on sparc? [04:51] No idea, I don't think hotplug has magic for that. [04:51] somehow it needs to I guess [04:51] Is the sbus driver sysfs enabled? [04:52] I could add that to initramfs-tools at least. [04:52] the debian installer handles it somehow [04:52] yes, it is [04:52] Debian's installer uses discover1 [04:52] discover, that's what I meant, not detect [04:52] Ah. 'k [04:53] I've just asked Colin to double check. [05:05] Yep. discover is only used for X autoconfiguration. [05:05] So possibly nothign does sbus detection at all. [05:06] Cookingup hotplug patches for sbus isn't hard, but I don't know that it could be done for breezy. [05:06] Best bet might be to do a plugin on sparc only. [05:06] If you can get a machine installed with Breezy (or if I can this weekend now that I have a nullmodem cable) [05:07] I can get the hack in for initramfs-tools to do any detection you need there. [05:10] something on debian loads hme(ethernet) and esp(scsi) modules [05:10] I don't know what it is === spayne [n=spayne@i-195-137-120-148.freedom2surf.net] has joined #ubuntu-kernel [05:11] i'm back to discuss my ndiswrapper problem :_ [05:11] i'm back to discuss my ndiswrapper problem :) === mkrufky [n=mk@68.160.103.77] has joined #ubuntu-kernel [05:13] Heya spayne, mkrufky [05:31] re [05:31] BenC: how is the installation going? [05:31] BenC: i am pretty sure you can manage to install ubuntu-standard [05:31] but not ubuntu-desktop [05:31] BenC: ping [05:33] well, it's headless, so desktop isn't needed anyway [05:33] spayne: ? [05:33] BenC: yeah so it's mine, but i still test full installability :) [05:34] BenC: can you talk about the ndiswrapper problem and see if it can be traced to something? [05:34] BenC: there is one package too much in ubuntu-desktop Depends: that makes it fail [05:34] fabbione: I read that as "instability" :) [05:34] spayne: sure, refresh me on what the issue is [05:35] BenC: lol :) [05:35] i think i will just build it manually and get over it [05:35] buildd is choaking on hoary-security [05:37] BenC: right, the bug related is http://bugzilla.ubuntu.com/show_bug.cgi?id=14147 [05:37] BenC: when i have installed and loaded up ndiswrapper, the kernel module isn't being loaded [05:37] it is weird [05:38] because if the device is plugged in (it is USB), the system hangs on 'Loading Modules' and/or 'Starting Hotplug' [05:38] ndiswrapper isn't loaded, or some other module [05:38] and when i remove the device [05:38] i get this error === doko [n=doko@dhcp-sn38-024.hrz.uni-oldenburg.de] has joined #ubuntu-kernel [05:38] Sep 22 21:21:14 amarillo kernel: [4294699.242000] ndiswrapper [05:38] (usb_reset_port:633): usb_reset_device() = -19 [05:38] Sep 22 21:21:14 amarillo kernel: [4294699.242000] ndiswrapper [05:38] (usb_reset_port:633): usb_reset_device() = -22 [05:38] Sep 22 21:21:14 amarillo kernel: [4294699.242000] ndiswrapper [05:38] (ndiswrapper_add_one_usb_dev:312): Windows driver couldn't initialize the device [05:38] (C0000001) [05:38] Sep 22 21:21:14 amarillo kernel: [4294699.242000] ndiswrapper: probe of 3-3:1.0 [05:38] failed with error -22 [05:39] which makes me think it is a USB problem [05:40] -22 == -EINVAL [05:40] what? [05:40] that's the error code it is returning [05:40] no idea what EINVAL means in this case [05:41] does it work under windows or any other linux dist [05:41] ? [05:41] yes [05:41] all other distros [05:41] be it suse, fedora [05:41] it works fine [05:41] SUSE 10 works the best and that has a 2.6.13 kernel [05:41] what version of the ndiswrapper module? [05:42] 1.1 - the standard one [05:43] is that the same version as in suse? [05:44] no [05:44] what version is in suse [05:44] give me a mo [05:45] 1.2-2 [05:45] We can't upgrade ndiswrapper at this point [05:45] why? [05:45] also, it works fine on my Dell laptop with a MiniPCI card [05:45] Because userspace needs to be kept in synchronisation with the kernel, and we're deep into freeze [05:46] BenC -- [05:46] applying patch external-drivers-net-sk98lin to ./ ... failed. [05:46] make[1] : *** [debian/monolith/patch-2.6.12-9.17] Error 1 [05:46] \o/ [05:46] mjg59: that's your patch failing to apply :) [05:46] neither of the sk98lin patches is working [05:46] hmmm [05:47] ok let's stop and focus one second... [05:47] zul's fails to build with some errors in skge.c [05:47] we can't keep uploading this way [05:47] I'm done uploading [05:47] Kamion is going with 9.14 for Colony 5 [05:47] did you upload .18 ? [05:47] ok [05:47] sk98lin will be an errata [05:47] i really don't think it's ndiswrapper [05:47] BenC: There's a one character fix to zul's patch [05:48]