=== T-Bone is now known as T-None === zul [~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 [01:06] echo echo === jbailey [~jbailey@CPE000ded9d787c-CM014260028338.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [03:23] its usually quiet tonight [03:24] yes, it is usually quiet ;-) [03:24] er...UNusually [03:33] Most of the usual culprits are testing the array stuff. I'm completing an install so that I can test md stuff. [03:33] I'm tackling a weird usb sound issues and trying to push through the rest of the Python 2.4 transition [03:33] issue, rather [03:34] what sound issue? [03:34] a disconnect issue that was fixed in -28 via the alsa patch === mdz [~mdz@ca-studio-bsr1o-251.vnnyca.adelphia.net] has joined #ubuntu-kernel [03:50] https://www.ubuntulinux.org/wiki/KernelBazNotes [03:52] register-archive, no? [03:52] (updated) [03:52] oh crap...i think i should go to beed soon [03:52] yes..bed.. [04:01] hmmm...lets see what rhel 4 have [04:13] later [06:06] morning [06:58] morning fabbione [06:59] hey lamont [07:54] fabbione: ARG. [07:54] svenl: ? [07:55] fabbione: i forgot to remove the non-powerpc configs yesterday, and it has not finished building :/ [07:55] ah ok [07:55] (its at power4-smp though) [07:55] fabbione: did you get my email ? [07:55] don't worry.. i am already building on our buildd [07:55] yes i did [07:55] fabbione: with my patch ? [07:55] patches even ? [07:55] i am buildining only the driver patches [07:55] not the initrd thingy [07:55] fabbione: why not both ? [07:56] because i need to test one thing at a time :-) [07:56] and see if one of them break === svenl don't get it what you all have against the mkinitrd support thingy ? [07:56] it makes it easier for me [07:56] fabbione: bah. [07:56] svenl: i will add it immediatly after this build cycle [07:56] it's not an upload [07:57] + everything is ccached [07:57] it takes only 30 minutes or so to build * [07:57] i promised to look at both of them [07:57] and i am doing it [07:57] so i don't see what the problem is [07:57] fabbione: they are two different things, the driver thingy will build you the modules, while the mkinitrd support thingy will copy stuff into the kernel-image. [07:57] i want a build box like yours :) [07:58] i know they are 2 different things. but since we have a policy to do atomic commits === svenl should maybe decide to use the augsbourg 8-way 1.6GHz power5 box for test build :) [07:58] i test A and commit A [07:58] fabbione: ok. [07:58] i add B -> test B -> commit B [07:58] it is seriously 100 times easier for me [07:59] svenl: that box is nothing special.. the ccache i have is [07:59] ok then, you have a fast box anyway. [07:59] you have a fast ccache ? [07:59] well i have a fast disk with the ccache already full [07:59] so it becomes fast [08:00] processor : 0 [08:00] cpu : PPC970FX, altivec supported [08:00] clock : 2000MHz [08:00] + 2GB of ram [08:00] and 500 or more GB of disk [08:03] fabbione: well, trick has 8GB of ramand who knows how much disk space. [08:04] fabbione: saddly we can't use it for upload-related development. [08:04] why not? [08:05] we have not full control over the box, so we cannot use it to build packages to upload. [08:05] i see... [08:05] suckage [08:05] as soon as i have time, i will use it to do ppc64 kernels though. [08:06] and hopefully i will get a virtual cpu on it someday. [08:06] you will need a libc6-64 and gcc-64 first [08:06] doko is working on those afaik [08:06] waldi already compiled a biarch toolchain two days ago. [08:06] i know doko had them for a while [08:07] not sure where tho [08:07] but i am sure they have been working together on it [08:07] fabbione: just the compiler, but with suse headers. [08:07] ok need to go, see you. [08:07] cya [08:17] fabbione: i wonder if it is possible to share a ccache cache setup with a chroot ? [08:21] yes it is [08:21] just use a dedicated partition that you mount under .ccache [08:21] and share it between chroots [08:21] it's up to you where you mount it [08:21] you have the env var to define it [08:22] just remember that you need to have the same UID under both the chroots [08:38] Changes by Sven Luther: [08:38] * Fix mv643xx driver for powerpc: [08:38] - Add patches powerpc-mv643xx-enet.dpatch and powerpc-mv643xx-eth-pegasos.dpatch. [08:38] - Obsolete patch marvell-pegasos-2.dpatch. [08:42] svenl: testing the initrd thingy... [08:42] how can i verify that it actually did something? [09:11] fabbione: do you have a /boot/vmlinuz-2.6.10-powerpc ? [09:11] svenl: i will in a short time.. [09:11] it's building with the mkwhatever patch [09:27] fabbione: what version of mkvmlkinuz do you have ? [09:31] the one you have sent to me via email === Seveas [~seveas@ksl403-uva-132.wireless.uva.nl] has joined #ubuntu-kernel [09:50] fabbione: the mkvmlinuz package, not the patch ? [09:51] no idea.... [09:51] let me check [09:52] 12 [09:52] but it is not installed [09:52] is it required as build-dep? [09:59] can you get mkvmlinuz 13 instead ? [09:59] altough not sure, it is more cosmetic. [09:59] no [09:59] we are in UVF [09:59] and that package is in main [09:59] mkvmlinuz 13 installs a postinst hook which asks the user. [09:59] fabbione: ok. [10:00] no problem, you don't really need mkvmlinuz, and it is just dead code right now. [10:00] ok [10:01] but make sure you mention this to me post-release as there will be some things to check, which may not be identic in debian and in ubuntu. [10:01] version 13, it asks for a debconf variable, if you want to use mkvmlinuz to create the vmlinuz kernel, nothing more. [10:01] (and adds support for 2.4 kernel, but you don't care about this). [10:02] svenl: after a release, a script is executed to sync * from unstable [10:02] so we will see what is in unstable after hoary [10:03] fabbione: version 13 is in testing. [10:10] fabbione: when will you have the package for me to test ? [10:14] svenl: it doesn't matter if it is in testing or not [10:14] we are too close to release to get a new upstream version [10:14] and i am quite busy right now.... [10:14] i will let you know when they are ready [10:16] fabbione: i said it is in testing, so we will have at least that in unstable after hoary. [10:17] http://people.ubuntu.com/~fabbione/ppc-diff [10:17] this is the output of debdiff [10:17] first deb is the old [10:17] second deb is the patched one [10:17] Mmm. [10:18] the script removes the .coff. [10:18] i don't like the fact that the /boot/vmlinux.coff-2.6.10-5-powerpc has been removed [10:18] I am not sure you want that, but i seriously doubt you are interested in the .coff. [10:18] and it installs in kernel-image [10:18] while it whould be linux-image [10:18] yep, but this is only because kernel-package is years outdated. [10:18] uh? [10:19] fabbione: you can generate the .coff with mkvmlinuz if you really want. [10:19] fabbione: the .coff is something that predates both quik and yaboot. [10:19] fabbione: nobody uses it anymore. [10:19] svenl: well i am not going to break ppc. we can remove it AFTER hoary [10:19] and like said, you can spare yourself the 1-2MB * 6 of space, since mkvmlinuz allows you to generate it. [10:20] fabbione: please ask Kamion first, but you can comment out one line in postinst-install.powerpc to keep it. [10:20] fabbione: but seriously, the .coff thingy is for a kernel which can be booted on oldworld pmacs through the serial console. [10:20] http://people.ubuntu.com/~fabbione/ [10:20] the images are there [10:20] fabbione: do you really envision a ubuntu/powerpc user doing that ? [10:20] svenl: listen.. i don't need to know all the history of ppc [10:20] and i don't really care [10:21] right now i have 2 problems: [10:21] 1) do not break the kernel [10:21] fabbione: ok, fine with you, just uncoment the line, ok ? [10:21] 2) fix major bug fixes [10:21] svenl: please just update the patch and send it to me [10:21] fabbione: it is just cruft because kenrel-package was never updates since ages. [10:21] oh perfect [10:21] ops [10:21] fabbione: come on. [10:21] ECHAN [10:21] that was for Jane [10:22] edit postinst-install.powerpc, and put a # before the line which says rm .... .coff. [10:22] fabbione: do you really want a patch from me ? [10:23] ok done [10:23] and changed the thing to linux- [10:23] fabbione: that said, mdz mentioned space problems on the powerpc isos, so i would check before removing this. [10:24] it may well be worth it to save almost 10MB of archive space, not sure. [10:24] but then you can always uncoment the line later on. [10:25] Victor Ferns <- right? [10:25] AMEN [10:25] i am too busy now... [10:25] fabbione: mmm, come to think of it, there is no chance of the .coff kernel being usefull, since it doesn't include the initrd,. [10:25] i need to do one thing at a time... [10:25] fabbione: ok. [10:25] fabbione: no problem. [10:25] fabbione: do you want me to mention this to Kamion/mdz and ask them if it is worthwhile ? [10:25] yes please [10:26] i am not the ppc porter [10:26] Ok, will do. [10:26] hehe. [10:26] i need people to work on it [10:26] without me [10:26] post-hoary i may take over that job for kernel stuff maybe. [10:42] fabbione: ok, installed the kernel and rebooting, i generated the mkvmlinuz kernel, will try to boot it directly. [10:43] fabbione: rebooted kernel worked. [10:43] mmm, i think we didn't enable the config option for the marvell gigabit ethernet driver :/ [10:44] yes we did [10:44] the driver is there [10:45] or did they changed again? [10:45] fabbione: well, i think it doesn't get autoloaded by hotplug. [10:45] mmm. [10:46] that is something that was fixed in the marvel-pegagos2.dpatch [10:46] fabbione: ah, will look at the patch, this means we need an additional patching step. [10:50] you sure ? I just looked at the marvel-pegasos2 dpatch from .27, and there is nothing in there relative to hotplug ? [10:51] no i am not sure [10:53] fabbione: can i access the kernel's changes list or something ? [10:53] svenl: there is no kernel changes list [10:53] we have 2 ml [10:53] kernel-team and kernel-bugs [10:54] fabbione: the one from the revision system repo (arch i think). [10:54] the latter collects all the stuff from the bug reports [10:54] there is none... [10:54] we don't have a ml with changes [10:54] there is only the repo [10:54] at the URL in /topic [10:54] fabbione: i mean the arch equivalent to svn log. [10:55] svenl: yes. you can access it doing a baz get of the archive in /topic [10:55] baz ? [10:56] or tla [10:56] doesn't matter [10:56] http://people.ubuntu.com/~lamont/Archives/kernel-team@ubuntu.com--2005/ this one ? [10:56] yes. register THAT archive [10:56] and get the branch pre29 in topic [11:07] mmm, need a crash course in tla :/ [11:13] damn, all that and the module doesn't seem to work as it should :/ [11:14] fabbione: no need to do an upload, i need to fix it first, and get the hotplug stuff worked out. [11:14] i am not going to upload until it is fixed [11:14] ok. [11:15] this will need more backport work, will not do this until this evening or tomorrow. [11:15] fabbione: what tla command do you use to do the checkout ? [11:16] did you register the archive first? [11:16] fabbione: i never used tla. [11:17] http://people.ubuntu.com/~fabbione/irclogs/ubuntu-kernel-2005-03-15.html [11:17] svenl: read there what lamont told to zul [11:18] you should be able to do s/baz/tla/ [11:24] fabbione: thanks. [11:24] fabbione: i am baffled, the driver works fine on 2.6.11, but fails in 2.6.10, need to look in details. [11:28] mmm, didn't work [11:28] tla make-archive tla-archive/luther@debian.org--2005 [11:29] usage: tla make-archive [options] [name] location [11:29] i am baffled. [11:29] later though, need to go now. [11:29] svenl: just look at tla register-archive and tla get [11:29] you don't need to create your own archive [11:29] otherwise install bazaar [11:29] and make your life simpler [11:30] $ tla register-archive http://people.ubuntu.com/~lamont/Archives/kernel-team@ubuntu.com--2005 [11:30] arch_archive_connect: attempt to connect to incompatible archive [11:30] than you need baz [11:30] oh well, will try again later. [11:30] baz is in ubuntu/hoary, right ? [11:30] yes afaik it is also in debian [11:30] E: Couldn't find package baz [11:30] package bazaar [11:30] no it doesn't seem to be. [11:30] oh, ok. [12:58] svenl: can you give me the changelog entry so i can commit the changes? [12:59] fabbione: added support files for mkvmlinuz [12:59] or do you want something more formal or more complete ? [12:59] maybe "mkvmlinuz support files added" even. [01:00] or mkvmlinuz support files added (Sven Luther) [01:00] don't know what the policy is. [01:00] + * Add support files for mkvmlinuz. [01:00] same as X [01:00] imperative form [01:00] or whatever is called [01:01] Ok. [01:01] what about attribution ? [01:01] you are already credited :-) [01:01] dude.... [01:01] ok, cool. [01:01] i am NOT a PUNKASS! [01:01] just curious. [01:01] Changes by Sven Luther: [01:01] * Fix mv643xx driver for powerpc: [01:01] - Add patches powerpc-mv643xx-enet.dpatch and powerpc-mv643xx-eth-pegasos.dpatch. [01:01] - Obsolete patch marvell-pegasos-2.dpatch. [01:01] * Add support files for mkvmlinuz. [01:01] + some spaces that IRC cuts away [01:02] hehe. [01:02] i do always respect credits [01:02] if i miss one it's for a mistake [01:02] quite verboser than the debian changelog format, but hey. [01:02] i am paranoid on changelog [01:03] well, the debian kernel changelog adopted the other form, because Jens vetoed the above (or rather d-i) format [01:04] claimed it did loose ordering of changes. [01:05] that's because he has no clue on how to use a RCS [01:05] fabbione: bah, he is inactive now anyway. [01:06] if you know how to use an RCS properly, there is no need to preserve the order of changes in the changelog [01:06] fabbione: provided all users who are interested need to know how to handle said RCS properly and have access too. [01:06] fabbione: i did the baz registering, how is the baz get done ? [01:07] baz get kernel-team@ubuntu.com--2005/kernel-debian--pre29--2.6.10 debian [01:07] it will get the archive for the debian/ dir [01:07] since it's the only one under rcs [01:08] thanks, altough i doubt the hotplug patch was ever included, i think. [01:09] afaik adding hotplug support is very simple [01:09] you need a table and one export of a function [01:09] pretty simple [01:09] just check how other drivers do that [01:09] yep. [01:10] fabbione: but the mv643xx_eth driver is not a pci driver, so ... [01:10] fabbione: how do i check the log or changeset of a file in baz ? [01:10] baz help [01:11] logs : list patch logs for a version in a project tree [01:11] baz logs --summary [01:12] for the short logs [01:17] WARNING: no rule found for checking signatures from kernel-team@ubuntu.com--2005 [01:17] Consider creating ~/.arch-params/signing/kernel-team@ubuntu.com--2005.check [01:17] or ~/.arch-params/signing/=default.check [01:17] PANIC: Unhandled arguments. Did you forget to use -r or -d? [01:17] its not helpfull, and baz help seems pretty useless to me :/ [01:18] mmm, === svenl guesses baz/tla don't allows to do per file logs like subversion does. [01:19] svenl: hold on.. you need to files to allow gpg checking of each commit [01:20] cat ~/.arch-params/signing/\=\=default.check [01:20] gpg --verify-files - [01:21] cat ~/.arch-params/signing/kernel-team@ubuntu.com--2005 [01:21] gpg --default-key 63549F8E --clearsign [01:21] of course you need to change the keyid [01:21] but you don't have commit access on our repo [01:22] so it's kinda pointless [01:22] yep. [01:22] let's forget it. [01:22] for now. === cc [~cc@byte.fedora] has joined #ubuntu-kernel [01:30] about hotplug, do you have a hint of the name of the structure or whatever ? [01:30] svenl: just check any other driver [01:30] it's a table with the pci ids [01:30] almost impossible to miss [01:34] something like this : [01:34] static struct pci_device_id rhine_pci_tbl[] = [01:34] { [01:34] {0x1106, 0x3043, PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, /* VT86C100A */ [01:34] {0x1106, 0x3065, PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, /* VT6102 */ [01:34] {0x1106, 0x3106, PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, /* 6105{,L,LOM} */ [01:34] {0x1106, 0x3053, PCI_ANY_ID, PCI_ANY_ID, 0, 0, }, /* VT6105M */ [01:34] { } /* terminate list */ [01:34] yes [01:34] }; [01:34] that is correct [01:34] i have such a one for doing a pci_dev_present thingy. [01:34] and there is an export of that table [01:35] so the hotplug is either : [01:35] MODULE_DEVICE_TABLE(pci, rhine_pci_tbl); [01:35] if you grep for rhine_pci_tbl [01:35] or : [01:35] static struct pci_driver rhine_driver = { [01:35] .name = DRV_NAME, [01:35] .id_table = rhine_pci_tbl, [01:35] .probe = rhine_init_one, [01:35] .remove = __devexit_p(rhine_remove_one), [01:35] #ifdef CONFIG_PM [01:35] yeah exactly [01:35] .suspend = rhine_suspend, [01:35] .resume = rhine_resume, [01:35] #endif /* CONFIG_PM */ [01:35] .driver = { [01:35] it depends how is the driver [01:35] .shutdown = rhine_shutdown, [01:35] } [01:36] }; [01:36] thanks will investigate. [01:36] i know both ways work [01:36] fabbione: its a none-pci-driver :/ [01:36] well.. i have no idea of non pci driver [01:36] but will ask around, i even know the guy to ask. [01:36] check how it is done in other non-pci drivers [01:36] but he is in awake-at-midnight timezone [01:38] Mmm, i think it is more complicated because the driver is no pci driver, but i will investigate. === smurfix [~smurf@smurfix.developer.debian] has joined #ubuntu-kernel === smurfix [~smurf@run.smurf.noris.de] has joined #ubuntu-kernel [02:10] svenl: What sort of driver is it? [02:11] The .id_table field is used by the kernel at driver load time - it goes through every device in the system, and calls .probe against each device that matches the contents of the table [02:11] MODULE_DEVICE_TABLE is used to export the information at build time [02:11] mjg59: the pegasos marvell northbridge gigabit ethernet driver. It is outside of the pci bus. [02:11] mjg59: the driver works, but i need to get hotplug to automatically load it. [02:11] Right. What sort of driver does it declare itself as being? [02:12] (As in, what's the type of the struct that contains the .probe and .remove functions) [02:15] svenl: Which driver is this? mv643xx? [02:16] If so, it doesn't seem to use the new driver model. There's no way to get it loaded via hotplug. [02:17] mv643xx_eth [02:18] Yeah. It looks like you're out of luck. [02:18] mjg59: huh ? [02:18] mjg59: well, we can adapt it, where can i find good info on the hotplug infrastructure ? [02:19] Documentation/driver-model/ [02:20] ok, thanks. === zul [~chuck@198.62.158.205] has joined #ubuntu-kernel [02:26] hey [02:50] morning [02:51] hey lamont how is it gonig? [02:51] morning lamont [02:51] hey fabbione [02:52] hi zul [02:54] we can be funny here.. right? [02:54] of course [02:54] most of us have a sense of humor except for lamont [02:54] and maybe jbailey === ..[topic/#ubuntu-kernel:fabbione] : Ubuntu kernel development discussion | http://www.ubuntulinux.org/wiki/KernelTeam | check the new abichecker code and GET familiar with it | http://people.ubuntu.com/~lamont/Archives/kernel-team@ubuntu.com--2005/ kernel-team@ubuntu.com--2005/kernel-debian--pre29--2.6.10 | Get it or http://www.getoffmynuts.com/gallery/images/asys.jpg [02:55] zul: I have no sense of humor. I'm Canadian. [02:55] zul: We have a sense of 'humour'. [02:56] jbailey: well excuse me i went to an american high school [02:56] fabbione: lol === lamont makes it back around to this channel, whaps zul with wet noodles [02:58] ENOTROUT === jbailey hands lamont a brick of tofu. [02:59] It's all I have to offer, sorry. [02:59] zul: my sense of humor is merely misunderstood. :-) [02:59] jbailey: yum [02:59] You can carve it to look like a trout. [03:00] heh [03:06] hey guys do you want to dump the 2.5 stuff in universe? [03:06] er...2.4 [03:07] http://www.getoffmynuts.com/gallery/images/chefno.jpg [03:07] aahah that looks so much like me [03:07] zul: It's probably a good idea. I know that the initrd-tools stuff that I've done hasn't been tested against 2.4 at all [03:07] fabbione: OooOooOoo! [03:07] heh its dom delouise [03:07] That could be the "Kernel Strike Force" page to go against the XSF one. =) === fabbione *coughs* [03:08] jbailey: i am kinda part of both.. you know? [03:08] but we can do something funny === dholbach [~daniel@td9091c17.pool.terralink.de] has joined #ubuntu-kernel [03:08] fabbione: Yes, well, we forgive you. =) [03:08]