=== Traxer|off [traxer@shell6.powershells.de] has joined #ubuntu-kernel [12:55] so, I have a unionfs patch that I need to get into breezy ASAP to unbreak LTSP [12:55] can someone explain the patch system to me? [12:57] mdz: you drop it into 00list-6.8, and give it the right name [12:58] lamont: any special procedures for a stolen-from-upstream patch to a third-party driver? [12:58] so that the patch goes away when the driver is updated? [12:58] the list there is somewhat ordered, pretty obvious where to add a new patch... the names are pretty straight forward as well [12:58] sth-* patches are stolen from head, and expect to go away soonish. [01:21] just tried dpatch-edit-patch and got: [01:21] Missing /tmp/dpep-ref.AieDu5/linux-source-2.6.12-2.6.12/debian/abi/2.6.12-6.7/abiname file. [01:21] make: *** [clean] Error 1 === mdz does it by hand === zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [04:18] morning === zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel === doko_ [~doko___@dsl-084-059-071-048.arcor-ip.net] has joined #ubuntu-kernel === jbailey [~jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-kernel === crimsun_ [~crimsun@sh.nu] has joined #ubuntu-kernel === Nafallo [nafallo@nafallo.user] has joined #ubuntu-kernel === fs [fs@213.178.77.98] has joined #ubuntu-kernel === Nafallo_ [nafallo@c-459571d5.07-44-73746f50.cust.bredbandsbolaget.se] has joined #ubuntu-kernel [06:35] mdz: there is already a 6.8 branch in baz [06:35] please use that one [06:35] is the unionfs patch part of a new upstream? or is it shipped separately? [06:36] * Temporarily disable fcpcmcia_cs; it needs more fiddling to work with [06:36] the new scheme [06:36] i guess that's because you used something like find . -name ".o" ? [08:23] fabbione: there is no new upstream; it is a bugfix from CVS [08:24] fabbione: the build rules for that module are confusing; I didn't immediately know how to handle it and I had to get the package out because all the other modules were broken [08:44] why is nvidia.ko so incredibly huge? [08:44] -rw-r--r-- 1 root root 4199482 2005-08-04 20:53 nvidia.ko [08:51] Because the binary from upstream isn't stripped? === chmj [~chmj@196.11.241.45] has joined #ubuntu-kernel [09:39] mdz: Because half their 3D stuff is done in kernel space, or something [09:51] mdz: for unionfs, please create a patch called external-fs_unionfs-fixups or similar to the others, and commit it to the baz archive. Ensure that the ABI doesn't change. [09:51] mdz: if it does, i will help you this afternoon at the meeting time. [09:52] mdz: (even if there are a bunch of convinient make targets for that ;)) [09:53] but please don't upload without updating the repo... [09:54] it will be messy to merge the bits later [09:54] and i don't mind if you commit directly in the playground (as topic) === zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel === doko_ [~doko___@dsl-084-059-071-048.arcor-ip.net] has joined #ubuntu-kernel [11:45] oh, fun. [11:45] Solaris/Xen === jbailey [~jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-kernel === chuck_ [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel === zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel === lamont [~lamont@15.238.5.30] has joined #ubuntu-kernel [04:08] hey lamont [04:08] morning [04:10] lamont: I have yet anoter build going to test this patch. *sigh* [04:10] With any luck... [04:11] jbailey: I'm doing a binary search, myself... [04:11] lamont: In this case, the patch I have already has what I want working. It just breaks the old initrd compatability case. This time I've just torn out all of their "new" version of doing it and put mjg59's old version in for that stuff. [04:11] So it really *ought* to just work this time. [04:12] I see [04:16] anyone got a 3c59x on a laptop? [04:19] Err.. You mean a card supported by it? [04:19] I might have one here. [04:20] 3com serial card in a laptop? that's kinda scary. :-) [04:21] yeah...i want to test a suspend patch that i found, apparently it breaks with acpi [04:23] lamont: nic :P [04:26] zul: doh. I knew that. [04:26] have several of them, but not in laptops [04:29] crappers.. [04:29] Anyone here know what Fabio's trick is to redline concordia when building kernel? =) [04:30] magic? [04:30] No, he said there's some env variable [04:30] We should standardize on something that means 'make -j 500'... Like DEB_BUILD_OPTIONS="fuckmeharder" === zul does his best doug henning impression === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-kernel [04:34] jbailey: ~bin/make ??? [04:34] heh doug henning rocks... [04:34] er, ~/bin/make, that is === fabbione yawns [04:43] CONCURRENCY_LEVEL=300 [04:43] man make-kpkg or read debian/rules [04:43] :) [04:43] zul: 3com cards are known to be broken on suspend/resume [04:44] where is the baz archive for the kernel? [04:44] mdz: hey.. /topic [04:44] mdz: i will be around for the time it's needed for the meeting [04:44] so i can help you on the fly if you need me to [04:44] mdz: you'll want the sftp://chinstrap/var/arch/... [04:44] sftp://people.ubuntu.com/home/lamont/public_html/Archives/kernel-team@ubuntu.com--2005 [04:45] my bad [04:46] I want sftp://p.eople or sftp://chinstrap? [04:47] people [04:48] "my bad" == people, not chinstrap [04:48] forget I even said anything about chinstrap this morning [04:48] the sftp url above is the one you want [04:48] and please oh please make sure your umask is correct... [04:49] yeah 002 === lamont curses abichecks, goes back to his binary searh [04:49] you love the abichecks! [04:49] fabbione: is there a trivial way to unpatch the build tree, short of 'clean'? [04:49] dilinger: yes we do [04:50] dilinger: not when I'm binary-searching and just want a *^(_ kernel/modules to throw against the wall. [04:50] once I find it, then I love them. [04:50] lamont: meh... apply the patches in 00list backwards? [04:50] hehe [04:50] or use dpatch unpatch target? [04:50] ah, very true [04:51] it's lovely how 20 torrents using no more than 50K each way can kill a 6Mb/1Mb adsl line [04:51] 50K aggregated [04:54] ccache is pure love though. [04:54] although support for java would be nice === lamont grumbles at the result of his binary search attempt... [05:29] (failed with only half of the patches, and then failed with the other half.) === lamont tries reverting -6.7 to remove all of the diffs from -4.4 through -6.6, to see if _that_ boots [05:47] fabbione: the current 2.6.12 kernel for amd64 tells me that frequency scaling is unsupported? [06:10] Well, on the upside my new patch DOES seem to notice that the DSDT is there for the initrd case. [06:10] It simply OOPSes the kernel now. === jbailey finds a bigger hammer. === lamont ships tactical nukes to jbailey, just in case. [06:38] DUDE. [06:38] I'm a Canadian [06:38] They don't let us have those. [06:39] Brass knuckles and crossbows, boy. [06:39] (Depends if the fight stays in the bar or gets out to the alleyway) [06:39] jbailey: tell 'em it's just a really F*NG big hammer. [06:39] the cross bow is in case they try to run away: [06:39] ? [06:41] fabbione: before the next abi bump, let me know... we could turn of ISA and EISA for ia64 - no such animals exist... [06:43] jbailey: poutine would be better though then you can suffocate it [06:44] lamont: Dude... Don't leave fingerprints. [06:44] It's hard to lift them off of an arrow. [06:44] And you want to wear gloves when using a crossbow anyway. [06:44] And no balistics to track with. [06:53] doko: i am in holidays.. fix your CPU [06:58] funny way of showing it [06:58] lamont: ISA isn't ISA slots - it's legacy interrupts [06:58] If ACPI is used to set up everything it's probably ok [06:59] Oh, and you could have ISA. Someone could add a cardbus bridge to a PCI slot and put a PCMCIA card in it [07:02] mjg59: you were just a bit faster than me ;) [07:02] lamont: PCMCIA can give you ISA as well as keyboard and other stuff [07:02] at least... it might have [07:02] so i would rather keep it there === doko_ [~doko___@dsl-084-059-081-178.arcor-ip.net] has joined #ubuntu-kernel [07:09] ok. sigh [07:20] lamont: trust me.. i really understand how you feel about that [07:20] i hate distro kernel configs === carl [~carl@217-162-189-134.dclient.hispeed.ch] has joined #ubuntu-kernel === carl [~carl@217-162-189-134.dclient.hispeed.ch] has left #ubuntu-kernel ["Verlassend"] [07:45] so, I need to get this unionfs patch in [07:45] it is not at all clear to me how to use the baz archive, there are dozens of branches, and I'm going to just upload a 6.8 unless someone wants to lend a hand with it [07:45] the unionfs bug is blocking all of my work [07:46] I already did a full test build of my tree [07:47] there is no mainline-2,6,12-6,8 branch [07:47] mdz: all of the 6.8 work is done on preX,8 [07:48] then it gets branched to mainline-2,6,12-6,8, and uploaded from there, and a preX,9 branch created [07:48] mdz: and the full test build was all 3 architectures, or just one or? [07:49] I can turn the crank if you want [07:49] lamont: just one, but the patch is very nearly a one-liner [07:49] you say "good", iz good for me. [07:49] your patch is committed to preX,8? [07:50] lamont: no, it is on its way to you via emil [07:50] email [07:50] it is an updated version of external-fs_unionfs-fixups.dpatch [07:50] ok. wanna toss a copy at lamont.jones@hp.com as well? [07:50] or however the _-_ goes [07:50] right [07:50] sent, thanks === lamont screams and pounds his head on the desk. [07:51] -6.7, with all of the patches between -4.4 and -6.6 removed, still fails to boot. [07:51] now for a serious diff. [07:52] mdz: mind if I stall on the upload for 20-40 minutes? or do you want it _NOW_ [07:52] ? [07:53] lamont: I would have uploaded it last night except that fabio yelled stop [07:53] but I can wait 30 minutes if necessary [07:54] ok. I'll make at most one more pass at figuring out ia64's issue, and then give up and call it -6.9 [07:54] and yes, I can build an ia64 kernel in about 20 minutes. not sure why the DC takes 2hrs/kernel [08:10] mdz: i am here... [08:10] mdz: commit the patch to pre6.8 [08:10] i have the time to look at it, upload and do all the baz dance [08:10] fabbione: patch is committed === lamont wants to commit one more hail mary patch for ia64, then he's ready [08:11] but first I need to make sure it's not an abi change [08:11] (and commit the ignore if it is...) :-) [08:12] lamont: i am not in a hurry :) [08:12] expect handoff within 15-20 minutes then - I'd rather have you do the baz dance [08:12] mdz: does the patch break the ABI? [08:12] or I could do it with you watching over my sholuder [08:12] fabbione: absolutely not [08:12] mdz: perfect... [08:13] lamont, mdz: we need to give jbailey sometime to integrate initramfs as default [08:13] it is very small and isolated and I need it ASAP [08:13] so i suggest we push these 2 changes now [08:13] yes [08:13] after that we will do a no-op upload (almost) to rebuild the kernel to use initramfs as default === lamont hurries up [08:14] (that requires a change in the B-D and one across the image Depends) [08:14] so we will have a very well defined break point [08:14] mdz: yes.. that's why i am still around :) [08:14] mdz: the patch is simple and looks good :) [08:15] lamont: i will wait for you than [08:15] mdz: is it ok for you to wait 20 minutes? [08:15] fabbione: lamont said he could do it [08:15] it will still take sometime to be built on the buildd and so on.. [08:15] you don't need to stay around if you have things to do [08:15] mdz: nah.. i am waiting dinner to be ready :) [08:15] and my mommy and daddy are there looking at it [08:16] fabbione: I'm down to the only diff between working and not-working being CONFIG_USB_BANDWIDTH=y( working) and =n(broke) [08:16] WTH!! [08:16] lamont: EH???? [08:16] yeah - exactly [08:16] that's impossible [08:16] exactly [08:17] I took -6.7, rolled all of the patches from -4.4 through -6.6 back, and it still doesn't boot. -4.4 with my -6.7 diff applied boots. [08:17] and I am very annoyed at it [08:17] it looks for all the world like device interrupts aren't working in the newer kernel [08:18] all of which may mean that I'm facing a bug that's present in both, but magically masked for whatever reason in the other [08:18] oh great... [08:18] would turning CONFIG_USB_BANDWIDTH back on for ia64 break abi"? [08:18] hmm it shouldn't [08:19] CONFIG_USB_BW is like 2 lines in the usb code [08:19] ok. I'll go ahead and commit that little hail mary bit, and then you can dance [08:19] just grep for it in drivers/ [08:19] yeah - nothing [08:19] or nearly nothing [08:19] lamont: works for me [08:19] but the weird is that USB doesn't do anything at boot.... [08:19] it must be a gtk bug [08:20] now.. how can i disable DPMS??? [08:20] given that it is blanking my screens every 2 seconds... [08:20] lamont: does the kernel work with USB_BW enabled? [08:21] I'll know that in about 20 minutes [08:21] did you check it, or you are assuming due to the diff? [08:21] ok [08:22] well, "hail mary" --> no real reason to expect it to fix things [08:22] but it's the only diff left. [08:34] fabbione: kernel build should be done shortly, I htink [08:35] lamont: i am going to eat something while you finish your tests. i will be back shortly. [08:35] ok === Seveaz [~seveas@seveas.demon.nl] has joined #ubuntu-kernel === lamont reboots, muttering something about what he may do if it boots, or doesn't boot [08:57] go crazy? [08:58] fabbione: go with what's committed currently - USB_BW didn't change a thing, just like we expected. [09:02] lamont: ok === lamont baz-undoes === fabbione starts the baz dance === lamont claps in time === fabbione uploads [09:16] mdz: enjoy :) [09:17] fabbione: thanks [09:17] mdz: no problem.. [09:18] we need to get rid of baz.... probably bzr could do... [09:18] or we should just move to git ;) [09:18] better idea.. let's go BITKEEPER :P === dilinger gives fabbione a gold star [09:20] dilinger: ehhehe [09:25] humpf.. damn connection is slow... === ..[topic/#ubuntu-kernel:fabbione] : Ubuntu kernel development discussion ONLY | http://www.ubuntulinux.org/wiki/KernelTeam | There are no kernel bugs.. only broken hardware | http://people.u.c/~lamont/Archives/kernel-team@ubuntu.com--2005/ playground: kernel-debian--preX,9--2.6.12 [09:29] katie should love me in a minute.. [09:29] and right before the end of the day [09:30] aiee...i just commited something to my preX,8...doh [09:30] zul: merge pre9 and branch again ;) [09:30] or branch and merge... [09:31] time to go offline [09:31] cya tomorrow guys [09:32] cya fabbione [09:33] toodles [10:43] http://people.ubuntu.com/~lamont/buildLogs/l/linux-source-2.6.12/2.6.12-6.8/linux-source-2.6.12_2.6.12-6.8_20050805-2048-i386-failed.gz [10:43] GRRR === lamont kicks vernadsky [10:49] what a strange error [10:50] oh, I understand [10:50] the chroot was broken, I guess [10:50] a package was half-installed and apt wanted to start over [10:50] or rather, reinst-required [10:51] right [10:51] given back