=== T-Bone wonders whether we have preliminary hppa installer he could test on his brand new hw === zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [03:06] hey [03:34] oh man. [03:34] even newer Nvidia drivers. [03:34] (http://www.nvidia.com/object/linux_display_ia32_1.0-7174.html) [03:35] (and http://www.nvidia.com/object/linux_display_amd64_1.0-7174.html) [03:36] fabio isnt going to like that :) [03:40] nope, he sure isn't. === Lathiat [~lathiat@gasp.bur.st] has joined #ubuntu-kernel [07:16] morning [07:16] morning [08:07] dilinger: ping? [08:07] pong [08:07] dilinger: you should probably get -33 and grab the security fixes [08:08] sorry i had no time to push them to you yesterday [08:08] they also assigned a CAN for your patch 143 [08:08] the sysfs signess whatever fix [08:08] thanks; already looked at -33 [08:09] zul found another 3 race condition fixes [08:09] for 2.6.11, anyways [08:09] he posted the bk commits in here [08:09] still need to deal w/ older kernels [08:09] oh? [08:09] just a sec.. i will forward the mail [08:10] on the way [08:10] one for sure needs to be applied [08:10] it's Herbert fix to his previous fix [08:11] they have not been classified as security yet [08:11] pitti is investigating the other 2 i think [08:12] ok === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-kernel === smurfix [~smurf@run.smurf.noris.de] has joined #ubuntu-kernel === smurfix [~smurf@smurfix.developer.debian] has joined #ubuntu-kernel [02:40] morning [05:22] ugh [05:22] -33 hppa64 doesn't boot :P [05:22] Unable to identify CD-ROM format. [05:22] cramfs: wrong magic [05:22] piKernel panic - not syncing: Attempted to kill init! === T-Bone reads backlog, notices XFS isn't tried, screams === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-kernel === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-kernel [05:26] we planning a -34? [05:26] hey lamont [05:26] morning t-bone [05:26] lamont: i've been flooding you in pv window ;} [05:26] yeah - saw it [05:27] get a good kernel [05:27] the good thing is that it's getting better [05:27] the bad thing is that your archive is corrupted [05:27] ?? [05:27] yeah, if you look a few lines above, i just tried -33 which doesn't know XFS, seemingly [05:28] /boot/config-2.6.10-2-64:# CONFIG_XFS_FS is not set [05:28] lamont: the Dubious version messages came from your archive [05:28] that'd be correct [05:28] lamont: that's bad, let's change that for -34 [05:28] lamont: your archive seems to contain packages taken from Debian. Their version numbers are newer than those in Ubuntu [05:28] (and yes, it's been disabled since the beginning of time...) [05:28] that's especially true for perl, which is incomplete, and thus not installable [05:29] lamont: please enable it in -34 ;] [05:29] "my archive" == ubuntu-hppa/tree or ubuntu-hppa/lamont? [05:29] /lamont [05:29] tree is just fine [05:29] except that it lacks up-to-date python2.4, which makes d-i unbuildable [05:30] that's why i'm trying to build python2.4 asap ;} === T-Bone larts lamont for bad archive management ;^) [05:30] ah, yeah. [05:30] that makes sense [05:30] heh.../me is working on the bk stuff === lamont makes a note to move all the stage[12] stuff out of /lamont [05:31] T-Bone: I only created that packages file because you wanted it... :)) [05:31] LOL [05:31] bad bad you :) [05:32] in any case, i've removed that archive from my sources.list and w-b merge, and i'm building all that's missing [05:32] xorg updated to -9 [05:32] currently building gcc [05:32] how do you create a new baz archive? [05:33] lamont: we're getting there. Down to 20% uncompiled [05:33] ;) [05:33] zul: bas make-archive -s ... [05:33] baz make-archive --help [05:33] okie dokie [05:33] most of it is kde shit not building because of yet uninstalable kdelibs4 [05:34] you mean kde poopie dont you? [05:34] i guess so ;) [05:35] lamont: i have a modified d-i source that i expect to install fine now. Need to build it (that requires python2.4 up-to-date) to confirm :) [05:54] T-Bone: xfs as a module, or =y? [05:54] module [05:55] and does it create any additional questions? [05:56] ? [05:56] nm - I'll go test it [05:56] lamont: look how it's done in debian or in other archs in ubuntu? === T-Bone needs to find a recent enough kernel that'd boot his j6k [05:58] lamont: btw, xresprobe builds fine now, kudos (as of 0.4.18) [05:58] # CONFIG_XFS_RT is not set [05:58] CONFIG_XFS_QUOTA=y [05:58] # CONFIG_XFS_SECURITY is not set [05:58] CONFIG_XFS_POSIX_ACL=y [05:58] that'll be a yes [05:58] yeah - daniel included my patch [05:59] ah, you mean "that" kind of questions :) [05:59] yeah [05:59] i thought you were talking about d-i :) [06:00] well, you still won't have XFS for install, since there's no module for it ... that's a kernel and d-i change [06:00] correct [06:00] guess I could have the kernel deliver the udeb [06:00] that'd be nice [06:01] i mean, let's do it as other archs do. My changes in d-i source were mostly copying what was done for ia64 [06:02] yeah === T-Bone wishes there were a way to tell the kernel build process "I only want that flavour please"... [06:04] T-Bone: there is but you have to modify the rules file [06:04] how big is that change? [06:04] one line [06:04] give it! :) [06:04] only if you say please [06:05] oh dear zul, please hand over your immense knowledge and kernel-build-fu to the poor beggar I am =] [06:05] better.. [06:05] ;} [06:05] comment out the flavours line in the rules file and replace it with flavours := hppa [06:05] zul: better watch out, or he'll find a tall iron building in new york and invoke you [06:06] heh [06:06] LOL [06:06] back to pope watch 2005 [06:08] yay...2.6.12-rc1-bk4 [06:09] 8GB RAM does *speed* up builds *alot* [06:10] its all in your head [06:10] i wonder :) [07:21] hmm [07:21] lamont, T-Bone: i have no issue in adding XFS for hppa in 34 [07:21] and yes there will be a -34 (99.9999%) [07:21] just be sure that it is propagated properly [07:21] because of me? :) [07:21] if you care [07:22] zul: yes. DIE! [07:22] :P [07:22] fabbione: cool [07:22] fabbione: do you know what's in ide-core-modules ? [07:22] lamont: do you have time to check something for me? [07:22] hmmm...how do i tell a user that he sucks...taticfully [07:22] we do not generate that udeb on hppa [07:22] fabbione: sure [07:22] T-Bone: you can check it yourself. install kernel-wedge and check [07:22] btw, mind if we dink around with what modules get built on hppa for -34? [07:23] lamont: can you try to test build kdebindings (last version) and see how much memory it sucks? [07:23] lamont: as i said about it is fine with me [07:23] drivers/ide/ide-core.o [07:23] zul: you either have to refine the meaning of 'sucks' to something objective, or accept the fact that he won't take it peacefully [07:23] something tells me we want that on hppa [07:23] just prebuild it to see if everything goes at it should [07:23] fabbione: cool [07:23] lamont: if you can run that build it would be really appreciated. here on sparc is sucking almost 1.2GB for one file [07:23] lamont: #8426 [07:24] g++ -DHAVE_CONFIG_H -I. -I/build/sparcbuildd/kdebindings-3.4.0/./smoke/kde -I../.. -I/build/sparcbuildd/kdebindings-3.4.0/./smoke/kde/.. -I/usr/include/kde -I/usr/share/qt3/include -I. -DQT_THREAD_SUPPORT -D_REENTRANT -Wnon-virtual-dtor -Wno-long-long -Wundef -ansi -D_XOPEN_SOURCE=500 -D_BSD_SOURCE -Wcast-align -Wconversion -Wchar-subscripts -Wall -W -Wpointer-arith -Wwrite-strings -DNDEBUG -DNO_DEBUG -O2 -g -Wall -O2 -Wformat-s [07:24] this one specifically [07:25] lamont: i am afraid of some memory leak somewhere, but the point is that it might pass our buildd, but not normal machines [07:25] i386 took 600+MB and 37 minutes [07:25] lamont: how shall i kill you? [07:25] i am talkign about RAM [07:25] WARNING: The following packages cannot be authenticated! [07:25] cdebconf-udeb archdetect anna casper-check cdebconf-priority [07:25] ... [07:25] it's up to 1.3GB now [07:25] fabbione: right [07:25] E: There are problems and -y was used without --force-yes === T-Bone hates that signed shit [07:25] T-Bone: add the key [07:25] and it's mvo/mdz you want to kill for it [07:25] lamont: how/where? [07:26] fabbione: right [07:26] lamont: my archive isn't signed [07:26] that's not the problem of the key [07:26] T-Bone: ah, well sign it [07:26] lamont: no rush, but i am just curios to understand wtf it is doing [07:26] it just can't cope with unsigned archive [07:26] lamont: how? [07:26] create a Release file, gpg --detach-sig Release; mv Release.sig Release.gpg [07:27] T-Bone: or use the workaround: [07:27] and make sure that the key you used is in /etc/apt/trusted.gpg in the chroot [07:27] cat chroot-hoary/etc/apt/apt.conf [07:27] APT::Get::AllowUnauthenticated "1"; [07:27] or do that [07:27] fabbione: ah! THANKS! [07:27] ;} [07:27] in /etc/apt/apt.conf [07:27] lamont: i'm 90% certain we need to produce ide-core-modules udeb btw [07:27] fabbione: I don't have kdebindings in my local mirror... and probably not the build-deps either [07:28] T-Bone: we do generate the modules, at least... :0) [07:28] lamont: perhaps you can run it at the DC? [07:28] fabbione: yeah - I'll do a build on concordia or some such [07:28] lamont: heh :) [07:28] lamont: danke [07:29] lamont: you see, in the ongoing port war, it is not acceptable that fabbione has d-i ready on sparc and we don't have it on hppa ;] [07:29] T-Bone: you will never have it for hoary anyway :) === fabbione rocks [07:30] fabbione: lol [07:30] fabbione: hoary releases next week? [07:30] sparc > hppa [07:30] fabbione, them's fighting words. [07:30] fabbione: not in number of built packages :) [07:30] T-Bone: how many buildds do you have? and with how many cpu's + gb of ram? [07:30] fabbione: i expect to have the installer ready by tomorrow, given a proper kernel [07:31] T-Bone: we had a booting miniiso before he did [07:31] fabbione: *LOADS* ;] [07:31] lamont: hehe cool! hppa > sparc ;) [07:31] mind you, you couldn't do anything with it. [07:31] T-Bone: than divide the num of your packages by (num of buildd * num cpu * Gb of RAM) [07:31] lol [07:31] T-Bone: and i will do the same [07:31] fabbione: I think we fixed xresprobe for you with 0.4.18 as well [07:31] the one with the higher num, wins [07:31] fabbione: i don't do maths :} [07:31] or was sparc already in the makefile the wrong (old) way? [07:32] Failed to fetch http://people.ubuntu.com/~lamont/ubuntu-hppa/tree/pool/main/c/console-data/console-keymaps-usb_2002.12.04dbs-48ubuntu5_all.udeb 404 Not Found [07:32] lamont: nope.. i did fix xresprobe for hppa === T-Bone curses lamont [07:32] lamont: and i did give you a kernel :) [07:32] when did you fix xresprobe for hppa? [07:32] because it hasn't built until 0.4.18, when daniels took my patch [07:32] lamont: i fixed xresprobe for hppa the same time i uploaded the fix for sparc [07:32] ftbfs [07:32] using the dummy_thing.x [07:33] well, you never told me === fabbione doesn't have a crystal ball yet [07:33] :) [07:33] ah no [07:33] xresprobe (0.4.12) hoary; urgency=low [07:33] * Fix ia64 and sparc port. [07:33] -- Fabio M. Di Nitto Sun, 21 Nov 2004 12:46:23 +0100 [07:33] that's because it's been ftbfs all along, and I'd sent daniels the patch to just make it use dummy _EVERYWHERE_ that it didn't use one of the other two, instead of hardcoding each *()^*_& architecture one at a time [07:33] i fixed ia64 [07:33] lamont: missing udebs are casper-check, di-utils, di-utils-reboot, di-utils-terminfo, hotplug-udeb, kickseed-common, localechooser, console-keymaps-at, console-keymaps-usb and initrd-kickseed [07:34] fabbione: ah, and I fixed everything else, and removed your ia64/sparc fix by subsuming it. [07:34] lamont: my bad.. it was ia64 [07:34] fabbione: heh [07:34] lamont: right [07:34] lamont: can you fix that in an easy way or should i build something? [07:34] well mdz is not around [07:34] i am going to help my wife to cook dinner === T-Bone begs lamont [07:35] later guys [07:35] fabbione: cya [07:35] toodles [07:36] linux-source-2.6.11_2.6.11-0.2: currently building [07:36] ARG! [07:36] T-Bone: bide [07:36] lamont: yeah sorry :} [07:37] lamont: so, what should I do? [07:37] T-Bone: those are all newer than what I have in my archive... [07:38] you may just need to build them [07:38] for quick-turn that is [07:38] what package builds them? [07:41] T-Bone: casper, debian-installer-utils, hotplug, kickseed, console-data [07:41] lamont: it's strange. The only package != universe I have left to be built are: [07:42] devel/debian-installer_20041227ubuntu23 [extra:uncompiled] [07:42] debian-installer/kbd-chooser_1.09ubuntu16 [-:uncompiled] [07:42] misc/archive-copier_0.1.3 [-:uncompiled] [07:42] unknown/palo-installer_0.0.6 [-:uncompiled] [07:42] which means that they are supposedly up to date in your archive... [07:42] it's possible that udebs aren't current in my archive.. [07:43] gah [07:43] i hate you :P [07:43] the target was 'debootstrapable' at the time, if you recall [07:43] lamont: problem is i don't have your changes to console-data to get it build [07:43] sec [07:44] not to mention that telling w-b to force rebuild those packages isn't gonna be cool [07:44] unless there is a magic command? === lamont pushes things around a bit [07:45] ah! Thanks you sooooooo much ;} [07:45] w-b --forget, w-b --take, echo "foo_ver hoary" > build/REDO [07:45] for each foo_ver [07:45] yumm [07:46] and finish that before the next time you run merge-quinn [07:46] more yumm [07:46] baz register-archive http://people.ubuntu.com/~lamont/Archives/lamont@ubuntu.com--hppa-2005 [07:46] is mirroring there now [07:46] lamont@ubuntu.com--hppa-2005/console-data--hppa--2002.12.04 [07:46] is the current console-data [07:47] baz get that I suppose? [07:47] on the bright side, once hoary releases, we won't have to keep merging the patch forward. [07:47] yeah [07:48] although it's not popultaed yet [07:48] No such package (console-data--hppa) [07:48] heh :P === lamont types his passphrase a few hundred times === T-Bone parses quinn-diff to figure out the version numbers of corresponding packages [07:49] i must really be damn willing to get that thing to work, cause it's kinda boring ;] [07:49] heh [07:49] you sound like my 9 year old. :-) [07:49] lol [07:49] i hate you soooooo much :} [07:49] "but it's _BOORIIINNNNGGGG" [07:50] that's the idea, yeah :) [07:50] T-Bone: do you want a lolly? === T-Bone ponders building missing udebs out of the buildds using his local sbuild [07:50] the only lolly I know is married [07:50] zul: you, i kill you, you! [07:50] T-Bone: that would be simplest === lamont starts an rsync of his local mirror up to p.u.c [07:51] lamont: definitely. Just need to find versions === T-Bone greps and greps more [07:53] hmm [07:53] i knew this wasn't that easy [07:53] /usr/bin/fakeroot debian/rules binary-arch [07:53] make: Nothing to be done for `binary-arch'. [07:53] dpkg-genchanges -B -mUbuntu hppa unofficial buildd [07:53] dpkg-genchanges: arch-specific upload - not including arch-independent packages [07:53] dpkg-genchanges: failure: cannot read files list file: No such file or directory [07:53] lamont: that's casper [07:54] ew [07:55] yeah, that's what I said too :P [07:55] isn't casper "all"? [07:55] Architecture: all === T-Bone giggles [07:56] same goes for hotplug [07:56] and kickseed [07:57] well, in that case, add that horrible hack to pull them from the real archive [07:57] i'm left with console data, and another tweak to sources.list.udeb.local to add something that will make it get all udebs [07:58] deb http://archive.ubuntu.com/ubuntu dists/hoary/main/debian-installer/binary-i386 ? [07:58] console-data source and debs in ubuntu-hppa/lamont if you want to go that route [07:59] much easier [08:00] lamont: 403 [08:01] grumble [08:01] i guess so :) [08:01] fixed [08:01] thx [08:03] btw, baz mirror of console-data done, working through ubuntu-meta now [08:03] ok [08:03] building console data [08:04] udebs were there too, you know... [08:04] yeah, but i can ;} [08:04] i don't trust that archive no more, you see... =] [08:04] mind you, it could be completely wrong, since it just clones the i386/ppc/whatever keymaps into hppa... [08:05] huhuhu [08:05] gonna be fun [08:05] funny part is that those debs you found aren't in changes files anywhere... [08:05] which is just plain wierd [08:05] for my current usage i don't need it anyway (serial console), but we'll have to look at that [08:05] yeah [08:06] my top priority is getting a netbootable installer image sufficient to net install machine using serial console, for next friday [08:06] once I have that, it'll be time to look at something more suitable :) [08:07] that's 99% of the way there, of course... [08:07] indeed [08:09] console-data successfully built === T-Bone tries again to build d-i [08:17] E: Malformed line 3 in source list /build/varenet/debian-installer-20041227ubuntu24/build/sources.list.udeb.local (dist parse) [08:17] shit [08:17] it doesn't like the hack [08:17] lamont: can you move the all.udebs in place or can you tell me how to do that? [08:19] fetch the packages.gz file from the right place, and lots of wgets [08:19] sigh [08:20] or edit the local sources list like it says [08:20] that's what i did [08:20] and it didn't like it [08:20] cat sources.list.udeb.local [08:20] deb http://archive.slashdirt.org/udebs ./ [08:20] deb http://people.ubuntu.com/~lamont/ubuntu-hppa/tree hoary main/debian-installer [08:20] deb http://archive.ubuntu.com/ubuntu dists/hoary/main/debian-installer/binary-i386 [08:21] sigh. [08:21] indeed :P [08:21] yeah - just grab the binary-i368/Packages.gz, grep for _all.udeb, and wget those lines [08:21] then apt-ftparchive that tree [08:21] will do [08:23] baz mirror complete === T-Bone attempts again to build d-i [08:27] lamont: btw, i'm quite worried at how gcc-3.4 build failed... [08:29] yay, all udebs were downloaded fine === T-Bone wanders away for a short while (dinner time) [09:00] meh.. [09:10] yatta, d-i built successfully [09:10] yatta? is that french french ;) [09:11] iieh [09:11] sole wa nihongo dessu ;] === T-Bone tries to figure zul's face, laughs ;] [09:13] mama toifea kusotare [09:14] lol [09:15] Freeing unused kernel memory: 316k freed [09:15] Setting up filesystem, pattempt to access beyond end of device [09:15] lram0: rw=0, want=16520, limit=16384 [09:15] eaKernel panic - not syncing: Attempted to kill init! [09:15] s e wait ... [09:15] Can't find /proc in /etc/fstab [09:16] hah hah === T-Bone looks at lamont, points at known problem, looks for a fix in d-i source === T-Bone larts zul ;] [09:16] its called karma === T-Bone notes he wants 64bit kernel, given the amount of RAM [09:20] T-Bone: huh? [09:21] T-Bone: sore wa nihongo desu [09:21] lamont: our initrd is too big [09:21] hrm.. fix that [09:21] how? [09:21] oh, there's the ramdisk thing [09:21] see what ia64 does for default args [09:22] boot with ramdisk=iirc [09:22] or ramdisk_size [09:22] or some such [09:22] that's what i'm trying [09:22] and it doesn't work [09:22] btw, u-h/lamont cleaned up to just have current versions of stuff in it [09:23] using ramdisk_size=16800 hangs the boot [09:23] isn't palo limited in the amount of memory it can use for loading kernel/initrd? === lamont would be asking Kamion about that, if it weren't 1 week before release [09:23] could be [09:23] SNAFU [09:25] nah, normally it works... [09:25] you could look at what debian/hppa does [09:26] debian/hppa has a much smaller initrd [09:26] kylem's last post about that was explicit [09:26] so what don't they include that we are? [09:26] pdc_iodc_bootin() died during seekread [09:26] it's not dead! [09:26] it just prompted me that [09:27] lamont: any clue what that means? [09:28] i suspect palo limitation wrt size of initial boot material === T-Bone wonders what's on the ramdisk [09:29] 14M build/udebs/ [09:42] lamont: where should i look at, supposing i want to remove a bunch of stuff from the initrd? === lamont shrugs === T-Bone hates initrds [09:44] 12M Mar 7 00:46 debian-di.img [09:44] i concur. [09:44] 13M Apr 2 21:13 ubuntu-di.img === T-Bone wonders [09:44] i hate cd images. === T-Bone removes ide support, hopes that'll makes things a bit better [09:47] let's remove usb as well === T-Bone contemplates removing cdrom support, all in all === T-Bone wonders how the ia64 installer works, given what he sees in the source tree, wonders if said installer *still* works [09:52] lessee... 112MB at 1KB/sec... that's gonna take a while [09:53] lamont: if i'm still able to do maths, if you're shapped at 56kbps, you should be able to do 7kBps (say round it at 5) :) [09:54] yeah - but there's another rsync, and this is unshaped (since I don't shape ssh) [09:54] tssks ;) [09:54] nah - interactive perfomance is needful, dammit === T-Bone votes for seemingly broken ia64 installer, adds random bits === lamont test-builds -34 on hppa [09:56] well, the config changes in -34 [09:57] lamont: i think there's a bunch of config options missing [09:57] and another bunch of d-i udebs missing [10:00]