[02:00] BenC: Apparently git isn't all setup yet, and CVS should be still considered canonical. [02:00] BenC: They'renot expecting to be over on git for at least another week. [02:00] who's that? [02:01] parisc? === zul [n=chuck@ubuntu/member/zul] has joined #ubuntu-kernel [02:26] BenC: its there now [02:31] BenC: Yes, sorry. =) [02:31] doh...mental note...dont put anything in /tmp [02:34] It's fine until you reboot. =) [02:36] duh... [02:36] :) [02:36] so when are you coming back? [02:37] Back where? [02:37] here === jbailey looks around. [02:37] I *am* here. [02:37] er...ottawa [02:38] Oh. Mmm. [02:38] Unlikely to be in 2005. [02:39] ah.. [02:39] true... [02:39] Maybe in January to hang out with Carlos for a weekend picking his brain on toolchain btis. [02:40] neato [03:14] finally, running 2.6.15-5.7 on all my machines except the a500 [03:15] even reverted to stable rt2500 for my mame cabinet [03:17] heh [04:11] infinity: ping [04:21] BenC: It's runnig on the ppc64 and the sparc boxes? [05:01] BenC : pong... I'm doing test builds of the vga16fb change... Can you hold off on 5.7 for a bit? [05:06] BenC : If you're itching to get it uploaded and go to bed, just do it, and I'll do a 5.8 later when I've tested this stuff more. [05:16] jbailey: running on my g4, not ppc64 though [05:16] jbailey: but definitely test it, since there were more ppc64 updates this time around [05:16] infinity: how soon before you can get the vga16fb patch ready? [05:17] infinity: also, what's the ETA on lrm? [05:17] infinity: update-initramfs and notifier changes are in 5.7, btw [06:40] mornig guys [06:40] hey fabbione [06:40] hey Ben [06:40] still up? [06:40] yeah, was hoping to catch adam again [06:41] I'd like to wait for the vga16fb patch before uploading -5.7, but I want to see how long he'll be with it [06:42] i think he is out for lunch [06:46] I've got -5.7 building cleanly on 5 architectures, I need to do an upload before something breaks it :) [06:46] aahha [06:46] bah crappp!!!!! [06:46] BenC : Just upload it, then. I may be up all night. :/ [06:47] I'm having a pretty miserable day of one step forward, two steps back. [06:47] so fltk doesn't export opengl stuff on sparc that makes openexr FTBFS that stalls all of KDE [06:47] GO KDE! [06:47] infinity: ok, I'll just do another upload when you get that patch ready [06:47] infinity: is there anything I can do to help with lrm, so we can get linux-meta updated? === _maydayjay_ [n=maydayja@ip101109.101.nas.net] has joined #ubuntu-kernel [06:49] BenC: nah.. [06:49] it's better infinity does LRM and get the blames :) [06:51] I'm more than willing to put the blame on infinity :) [06:51] even if I break it :) [06:53] ehehhe === _maydayjay_ [n=maydayja@ip101109.101.nas.net] has joined #ubuntu-kernel === ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | New git tree for dapper: https://wiki.ubuntu.com/KernelGitGuide | 2.6.15-5.7 uploaded (The "Five Arches" release) [07:00] lol [07:01] congratulation dude [07:01] don't congratulate me till the buildd's agree with my assessment :) [07:02] this time though, I did all my builds through to udeb builds, so I should be good to go [07:03] ehhe [07:03] ok fltk1.1 fixed [07:04] kde will unleash soon [07:04] good night ben [07:04] cya later [07:04] good night [07:25] BenC : No, I just need to figure out some minor details. [07:25] dilinger: ping? === chmj [n=chmj@wbs-146-184-43.telkomadsl.co.za] has joined #ubuntu-kernel === jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-kernel [01:47] BenC: Checked ppc64, it still fails with the same rtc failure even though you've built it in. === zul [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [02:17] heylo === oldguy [n=cliebow@esmith.sandbox.ellsworth-hs.ellsworth.k12.me.us] has joined #ubuntu-kernel [02:25] quick ot ok/ [02:27] oldguy: Dad? [02:27] =) [02:28] yeah i m a dad..but got three girls [02:28] I have long hair, but that's as close as I get, sorry. [02:29] im trying to build a 2/6 kernel for netvista and ltsp-4.2..i built an uncompressed kerenl and catted the initramfs onto it..but get unsupported file type.. [02:30] so if vmlinux is unsupported..and vmlinuz is unsupported..what else is there? === chrisle [n=chris@p54BC83AD.dip0.t-ipconnect.de] has joined #ubuntu-kernel === CataEnry [n=Enrico@host204-30.pool8248.interbusiness.it] has joined #ubuntu-kernel [02:38] hi all [02:38] hi === chrisle [n=chris@p54BC83AD.dip0.t-ipconnect.de] has left #ubuntu-kernel ["...] === chrisle [n=chris@p54BC83AD.dip0.t-ipconnect.de] has joined #ubuntu-kernel === mgalvin [n=mgalvin@ubuntu/member/mgalvin] has joined #ubuntu-kernel [03:06] BenC, fabbione: What's the word on Xen on Breezy? [03:07] oldguy: You can't cat the initramfs onto a kernel. You have to embed it in the data segment. [03:07] oldguy: Or you can pass it in using the initrd mechanism. [03:14] jbailey: Xen has some kind of .12 support afaik.. but that's it [03:14] jbailey: last time i did try a merge they were at .11 [03:14] and we were shipping .12 [03:14] But it's nothing we support natively. [03:14] nope [03:14] 'kay, thanks, [03:25] fedora's .14 based kernel has patches [03:27] Ah, cool. [03:27] kick ass, the big three have successful builds for -5.7 [03:27] jbailey: feel free to test -5.7 on ppc64 when you get a chance [03:27] It's more someone I have looking for us to support Xen. [03:27] BenC: See my comment to you in this channel from a little under 2 hours ago. =) [03:28] jbailey: I looked into it, but the patches are so big and so invasive, I didn't see the point trying at this point [03:28] ah [03:28] hmm [03:28] jbailey: let me see if there's a kernel command line option that can disable rtc or something [03:29] if we really really want Xen we can pull it in from their development branch [03:29] BenC: 'k [03:29] fabbione: is there a Xen git? [03:29] IIRC, they are going to try to get it into 2.6.16 [03:29] BenC: there is an hg tree [03:29] the tool that's similar to git [03:29] it's also on kernel.org afaik [03:30] fabbione: -5.7 hit your sparc buildd yet? [03:30] http://www.kernel.org/hg/ [03:30] BenC: it's building as we speak [03:32] jbailey: can you past the URL to that oops again, please? [03:32] http://www.cl.cam.ac.uk/Research/SRG/netos/xen/downloads.html [03:32] The public Mercurial repository for Xen, the modified Linux, and the domain control tools is hosted on http://xenbits.xensource.com, where you can browse the source and inspect the [03:33] 15 hours ago: The device sharing check races when more than one file backed vbd is [03:33] author: rread@ubuntu.eng.hq.xensource.com [03:34] *perhaps* we can try to get in touch with that guy :) [03:35] BenC: Hmm, grep isn't turning it up in my logs. [03:35] let me check firefox history, what host name would it be? [03:36] That's what I'm trying to remember. =) [03:36] lol [03:36] Something ending in .nl =) [03:37] grep pastebin.ubuntulinux.nl * | grep jbailey [03:37] isn't giving me anything. hmm [03:37] No, right! [03:37] I didn't put it in there. [03:37] I took a picture instead. [03:38] fabbione: pong [03:38] http://people.ubuntu.com/~jbailey/P1010042.JPG [03:38] thanks [03:38] BenC: If you want I can update it for -5.7 [03:38] yeah, that'd be nice. see if you can get the whole screen aswell [03:40] dilinger: hey.. thanks i did manage without :) [03:43] BenC : Okay, LRM ABI switches are now a one-button affair, much like linux-meta, but MUCH SCARIER under the hood. [03:43] BenC : So, once this thing &#@$ing builds for me, your minor maintenance tasks with it should be a no-brainer. [03:43] if I ever have to pop the hood, I'll be sure to take it to the shop(you) rather than work on it myself :) === infinity kisses hardcoded kernel versoins and annoying *_minor stuff goodbye. [03:45] Of course, now I have an irritating QT build issue to deal with. === BenC adds ia64 to the successful builds for -5.7 [03:48] About time. [03:49] so far everything has built fine [03:49] I expect sparc to build, but I know hppa will fail === infinity kicks weddell and screams at it to be faster. [03:50] BenC: sparc is at 64-smp (half way) [03:50] it had been so long since I booted my i2k this weekend, I actually had to smack it on the side of the case to get it to power on [03:50] it won't take much much longer [03:51] fabbione: sweet [03:51] i am falling asleep and partman-auto-lvm on ppc is winning 2:0 === fabbione starts to get pissed off [03:51] time for coffe and smoking bof [03:56] opinions...in linux-meta, because {686,k7}-smp are going away, should I make the linux-{image,headers}-{686,k7}-smp meta packages depend on linux-{image,headers}-{686,k7} meta packages or make them depend on the kernel-abi-version packages directly (sans the -smp)? [03:57] Make them depend on the other meta packages, so we can eventually remove the SMP packages. [03:57] that's what I was thinking [03:57] (Assuming we want to down the road) === lamont [n=lamont@mib.fc.hp.com] has joined #ubuntu-kernel [03:58] lamont: ia64 built :) [03:58] BenC: woot! [03:58] and hppa is still git-ftbfs? [03:58] yeah [03:59] if we had 64-bit hppa64 userspace, it wouldn't be a problem :) [04:00] well, yeah [04:01] lamont: willy said to keep using CVS for at least another week. [04:01] ah, I see. [04:04] probably take a week to get a decent patch out of CVS anyway :) [04:05] anyone packaged git 0.99.9x for ubuntu yet, or even debian? === mgalvin [n=mgalvin@host-66-202-95-170.spr.choiceone.net] has joined #ubuntu-kernel [04:06] jbailey: thanks..gonna try an experiment [04:06] debian has it...we should grab that to dapper [04:08] BenC: i think we already sync that automatically [04:08] universe? [04:09] yep, there it is, thanks [04:12] jbailey: can you disable the hwclock problem (mv it out of the way, and symlink to /bin/true, e.g.) and see where you get after that? [04:12] BenC: You scare me. [04:12] But okay. =) [04:12] disable hwclock program I meant [04:12] Is that really the only thing that will cause an rtc event? [04:13] Or should I avoid using this system for much after that? [04:13] rootcd /opt/lytsp/i386-4.2 [04:13] nothing else should [04:13] 'k === jbailey tries [04:13] everything else should use the system clock rather than the rtc directly [04:15] (should) === jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-kernel [04:20] odd that his system seems to oops/freeze on sig 7 from I/O accesses, I would have thought that hw traps would recover from that [04:20] jbailey: running? [04:20] $ uname -a [04:20] Linux starshine 2.6.15-5-powerpc64-smp #1 SMP Tue Nov 29 07:20:51 UTC 2005 ppc64 GNU/Linux [04:20] sweet [04:20] ready to attempt to crash it? :) [04:20] cat /proc/drivers/rtc [04:20] The only observation so far is that power mgmt is also a bit broken. [04:21] The fans are slowly speeding up. [04:21] odd [04:21] Should I be on the console for that? =) [04:21] yeah [04:21] NO such file or directory [04:22] mm [04:22] driver singular. === oldguy [n=cliebow@esmith.sandbox.ellsworth-hs.ellsworth.k12.me.us] has left #ubuntu-kernel ["Leaving"] [04:22] BenC: I got an oops, but no crash. [04:22] well, it's a start [04:22] can you email me the oops? [04:22] Did you actually get any output? :) [04:23] yeah, did any output come out, or did the cat process lockup? [04:23] BenC: http://paste.ubuntulinux.nl/5190 [04:23] well, it would have to get to the end of the proc function before the buffer was returned to the user [04:23] jbailey: thanks [04:24] (Recovered from dmesg so I can just paste) [04:24] that's not what I expected [04:25] I was expecting to see rtc_get_rtc_time() again [04:26] heh, that can't be good :) === jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-kernel [04:26] jbailey: thanks, I'll be able to do more with this dmesg now [04:27] maybe I'll have something you can test later, if you have time [04:27] Apparently doing it a second time to see if there was any output crashes the machine. =) [04:27] at the very least, a little sprinkling of printk's to see where it actually oopses [04:27] BenC: Cool. I'll try to run anything I need long term in screen sessions on other machines. =) [04:37] infinity: At some point when you're bored... ;) [04:50] ? [04:50] Not that that point is now, but I may want to plan for the future. [04:51] Is usplash working for everyone now? [04:51] And what magic does it need in its postinst? [04:52] a) Don't know [04:52] b) update-initramfs -u [05:00] infinity: Is the 640x400 patch submitted yet? [05:00] I want to shift to 640x400 artwork in the next usplash upload [05:14] Too focussed on making LRM work. :/ [05:14] If yo uhave a tested patch, give it to BenC, otherwise I'll do mine tomorrowish. [05:14] infinity: I have instruction here on making kdump work to store OOPS information when it happens. [05:15] infinity: I collected it at OLS last year, and dug it up a couple days ago. [05:15] Though, you should be able to switch the artwork anyway, since artwork that's smaller than the fb isn't a problem. [05:15] Yeah [05:15] (As evidenced by the tiny artwork in the middle of my massive vesafb) [05:16] usplash looks tiny on my readonfb at 1024x768 [05:16] (What will happen with artwork that's bigger than the FB?... For people upgrading?) [05:16] anyway to scale usplash on higher resolutions? :) [05:16] BenC : Did you see it on my 1400x1050? [05:17] nah, hehe [05:17] Speaking of seeing things on my laptop, did you finally finish watching Family Guy and regain productivity? :) [05:17] I have about 10 episodes left to watch :) === mjg59 uploads new usplash [05:18] Also: usplash needs to check if the display is capable of 256 colours [05:18] mjg59 : I don't suppose that one daemonises (and removes the sleep from the init)? [05:18] Nope [05:18] I'll do that next time [05:18] Kay. I'll do that the first time I touch it, then. [05:18] Can you file a bug? [05:18] Unless you're on a warpath. :) [05:18] Well, I'm in bed, ill [05:19] Which, obviously, leads to hacking graphical boot thingees. === lucasvo [n=lucasvo@www.wservices.ch] has joined #ubuntu-kernel [05:39] hi [05:41] I would like to make a clustering system with upstream compatible Kernel for terminal server. What do you suggest? [05:41] I have found scary ways of helping vesafb survive suspend/resume [05:42] lucasvo: I don't undetstand your question [05:42] terminal server and clustering conflict in my brain [05:42] BenC: I would like to develop clustering which may be included by ubuntu [05:43] what sort of clustering? [05:43] mjg59 : It always has done for me anyway. [05:43] if one uses powerful ltsp clients, one could use them for clustering [05:43] BenC: like openmosix [05:43] infinity: Yeah, it doesn't here [05:43] mjg59: FYI, the fb-no-write patch is in -5.7 [05:44] BenC : LRM has turned into a mauch larger headache than I'd originally hoped. I'm going to go catch a 4 hour nap (It's 4am now) and finish up in the morning. :/ [05:44] lucasvo: oh, you lost me, I don't do much with that sort of thing [05:44] BenC: Thanks [05:44] BenC: ok [05:44] infinity: ok [05:44] lucasvo: but I still don't know what you are asking..."what do you suggest" is very vague :) [05:45] BenC: I need a clustering system which is accepted by ubuntu policy [05:46] we have ocfs2, but I'm not sure if that suits your needs, since it's only a clustering filesystem [05:46] BenC: exactly, I need RAM/CPU sharing [05:46] lucasvo: we don't have that solution yet [05:46] and "accepted by ubuntu policy" is something you'll have to gain somewhere down the line, it's not something that can be decided before hand [05:46] lucasvo: and probably won't [05:47] lucasvo: what kind of clustering exactly are you looking for? [05:47] fabbione: what is the clustering thing that was talked about at UBZ? [05:48] lucasvo: something like openmosix or a batch/queue/job scheduler? [05:48] some IBM thing [05:48] fabbione: yes, I want to make one for ubuntu [05:48] fabbione: yes [05:48] lucasvo: yes it's not an answer to what i did ask :). [05:48] there's another guy wanting to do the same thing [05:48] and pretty much, he is having to do his work outside of ubuntu, in the hopes that he can prove its viability for dapper+1 [05:48] fabbione: hm I didn't see the or [05:49] fabbione: I mean something like openmosix [05:49] BenC: i don't remember talkign about IBM solutions really.. but we have something on the boilerplate [05:49] lucasvo: no, we won't have openmosix or similar solution for some time [05:49] fabbione: maybe it was HP [05:50] BenC: yeah HP [05:50] openssi [05:50] lucasvo: email -devel about it and see if you can catch the attention of other people (and there are some) looking to do the same [05:50] fabbione: what I would like to do is to include a clustering system into ubuntu which is accepted by upstream [05:50] yeah, that's it [05:50] BenC: ubuntu-devel? [05:50] lucasvo: https://wiki.ubuntu.com/UbuntuClusters [05:51] lucasvo: most of solutions like openmosix are for 2.4 kernels only [05:51] lucasvo: yeah [05:51] and we don't support 2.4 [05:51] porting to 2.6 is on the way but still at a too early stage to be even considered [05:51] fabbione: there is some 2.6 code at openmosix but it is not accepted by Linus [05:51] as we discussed at the last Developers Summit, a group will have to develop this outside of ubuntu proper, and then seek inclusion, so I suggest getting on the bandwagon with people who want to do the same to avoid extra work and conflicts [05:52] lucasvo: because the userland is broken. [05:52] and the patch for 2.6 probably sucks [05:52] BenC: exactly [05:52] lucasvo: accepted by upstream as in linus? [05:53] don't see that happening with many clustering systems [05:53] I mean this use case: Hospital delta is deploying hundreds of thin clients with Ubuntu and LTSP, and wants complete failover for the terminal services. [05:54] BenC: yes [05:54] it will get into ubuntu before it gets into kernel proper :) [05:54] lucasvo: that's not an openmosix cluster [05:54] why the criteria for inclusion with linus? [05:54] that's a HA cluster [05:55] fabbione: yes [05:56] fabbione: but I actually want HPC [05:56] lucasvo: we don't have it and we won't for a while :) same answer as before. ;) [05:57] lucasvo: we might be able to provide slurm (job queue scheduler) [05:57] but that's it for dapper [05:57] fabbione: I am not talking about dapper more about dapper+ 1 or 2 [05:58] fabbione: and I would like to develop it, I don't want out of the box solution [05:58] lucasvo: from here to dapper+1 there are at least another 10 months.. perhaps by that time there will be available solution.. [05:58] who knows.. === mgalvin [n=mgalvin@ubuntu/member/mgalvin] has joined #ubuntu-kernel [06:08] hola fabbione [06:08] hey zul [06:09] yo zul [06:10] hey benco [06:11] jbailey: ping [06:17] ugh..this is an ugly website [06:17] http://wwwredesign.gentoo.org/ [06:19] BenC: sparc is at dpkg-deb now :) [06:20] fabbione: sparc success will finally mean an upload went as planned :) [06:20] ehehhe [06:21] goal for -6 is to disable all the drivers that have missing symbols (mostly smp and 64-bit incompatible drivers) [06:22] yeah [06:22] i saw a bunch [06:24] wow, linus tagged 2.6.15-rc3 just a mere 29 minutes after I tagged 2.6.15-5.7 [06:24] my timing is improving :) [06:25] heh [06:25] ehhe [06:25] slacker [06:25] :) [06:32] BenC: i wrote a patch for the atkbd last night its in my git archive [06:32] zul: patch to do what? [06:32] remove that unknown key bullshit [06:37] fabbione: http://zulinux.homelinux.net/cgi-bin/gitweb.cgi?p=ubuntu-2.6.git;a=commit;h=1ec11f9bc0963516a9b944e43fb89fbf5545a27b [06:40] silly printk, warnings are for developers [06:41] bah it's just a warning [06:42] linux-source-2.6.15_2.6.15-5.7: successful [06:42] but it causes uneeded bug reports === fabbione ^5s BenC [06:42] sweet [06:43] cool, latest git has --no-merge for git-pull [06:43] so you can inspect before merging and committing [06:44] eheh [06:44] i didn't play with git that much [06:44] i rely mostly on what you tell me [06:44] + my ppc is down until i can unfuck partman-auto-lvm [06:48] fabbione: i know but a user complained about it and its annoying the fuck out of me on my laptop [06:48] BenC: [06:48] fabbione willy: i am cool with that.. ben is merging rc3 and i was wondering when we could pull from you.. i expect lamont wants that ;) [06:48] willy fabbione: keep pulling from cvs [06:48] ^^ hppa tree [06:49] fabbione: not only that, but even if we switch it to KERN_INFO, it will still cause uneeded fillage of kern.log (and clearing of the dmesg buffer) [06:49] fabbione: I'll leave it up to lamont to get me a diff from CVS [06:50] BenC: is there actually anything that use that output? iirc we were using it to track kernel -> X -> gnome keyboard issues [06:50] BenC: didn't you pull from hppa git once? [06:50] fabbione: yeah [06:51] ok [06:51] BenC: once they finish the rc3 merge, I'll gen a diff [06:51] not sure about tracking the keys [06:51] lamont: better if you push them to move to git :P [06:52] yeah, there's a freaking CVS tool fot git anyway :) [06:52] fabbione: the trick is to help them get the infrastructure git-happy, and then it'll happen... it's not that they don't want to move. [06:52] BenC: its the autobuild stuff, and such [06:52] ah [06:52] bunch of home-grown tools to do thing [06:52] s [06:52] lamont: yeah i understand that... [06:53] I'm a little leary of taking a cold monolithic patch for hppa, if they will get around to being in git real soon [06:53] that's exactly what i was thinking [06:53] maybe we can do what fabbione and I discussed and just run patch/unpatch for that one diff [06:53] but you might endup duplicating the merge work [06:54] or I might be off by a line or two and cause uneeded deltas in our tree [06:54] since they do it and you have to do it to pull in rc3 [06:54] perhaps talk with them and see how to handle it best? [06:55] gah. git-core goes through 0.99.9i, and then it's git_0.99.9j - wish they'd make up their mind [06:56] lamont: apt-get install git-core (from dapper universe) [06:58] fabbione: bah, git-core is not built for sparc yet! :) === BenC [n=bcollins@richmond-209-163-125-167.dynamic-dialup.coretel.net] has joined #ubuntu-kernel [06:58] BenC: hmmm probably [06:59] git-core: [06:59] Package : git-core [06:59] Version : 0.99.9j-1 [06:59] State : Needs-Build [06:59] yeah === BenC tries to remember that Alt+F4 in X will not get him to VT4 [06:59] ahhahahahah [07:00] BenC: the universe buildd is almost at the end of gcc-snapshot [07:00] that's probably why it didn't built yet [07:00] and i did unleash the kde mess just today [07:01] so the main buildd is like utterly busy [07:01] fabbione: FYI, if you have trouble finding an affordable ultra for the DC, I know where you can get UE 4500's for pretty cheap [07:02] BenC: you should tell that to elmo/znarl.. but afaik they are looking into new hw [07:02] it needs to be as small as possible (size wise) [07:02] because we pay hosting by size [07:02] but i am not sure 100% [07:02] things keep changing on a daily base [07:03] yeah, the 4500 isn't small, but it's cheap and powerful [07:03] yeah i know === jane_ [n=JaneW@wbs-146-174-135.telkomadsl.co.za] has joined #ubuntu-kernel [07:06] ok time to stop for today [07:06] or better [07:06] start cooking some real food === lamont builds -9k on breezy [07:16] git push ssh://rookery.ubuntu.com/home/lamont/public_html/archives/ubuntu-2.6.git [07:16] bash: git-receive-pack: command not found === lamont grunbles [07:16] and grumbles, too === _maydayjay_ [n=maydayja@gimel.nas.net] has joined #ubuntu-kernel [07:16] BenC: would you be so kind as to send RT a request to install git on rookery? [07:18] meanwhile, I've updated ~/.bashrc [07:18] on rookery, you can point to my git bin's [07:18] only if .bashrc says to === lamont has a copy of git in ~/bin [07:27]