[12:10] few, past midnight, high time to head home [12:12] sabdfl gave the greenlight for the rh cluster stuff [12:13] gfs & co will be part of the official kernel [12:13] i got the patches in already, but i need to clean them up === atc_work [~stephen@cossox2-ext.cos.agilent.com] has joined #ubuntu-kernel [12:59] and I'm back [12:59] feh [01:00] lamont: machines are all ready, the rest is really up to you now... [01:12] T-Bone: cool. working on taxes tonight/tomorrow night. Then I'll get the hoary archive all happy, and then we get to bootstrap hppa's breezy into gcc-4.0 toolchain happiness [01:13] schweet [01:13] given everything, I expect I'll poke you sometime friday night/saturday ish, but I'll have a better schedule idea come friday morning... [01:13] btw, do you know where the ia64 archive lives? I'd like to install/upgrade some machines [01:13] ok fine :) === atc_work [~stephen@cossox2-ext.cos.agilent.com] has left #ubuntu-kernel [] === T-Bone supposes the ia64 archive currently doesn't have a place to live, figures he needs to wait === T-Bone is now known as T-None [02:37] T-None: the ia64 archive lives the same place as the sparc archive. :-( [04:59] damn [04:59] Subject: [patch 001/198] mmtimer build fix [04:59] lotta patches.. === zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel [05:10] heylo [05:18] lamont: are you around baz question for you [05:20] well, for you? sure. [05:21] :-) [05:21] cool...i can feel the love [05:21] ok..say that i am starting a new repo and i do all the crap that you usually do [05:22] including remembering --listing if it's a web-accessible repo [05:22] and always --signed [05:22] so everynight i do a new baz get kernel-team@ubuntu.com--2005 blah blah [05:22] ok. you could just baz update in a tree that is already taht version... [05:22] but when i branch zulcss@gmail.com--2005 it looks like it clobers my stuff [05:23] clobbers how? [05:23] baz branch is going to create a new branch with the given name... [05:23] which could overwrite what you've done... [05:23] ah... [05:23] you really want baz switch, and baz merge type stuff [05:24] that is, in a zulcss@... repo, baz merge kernel-team@ubunut.com--2005/kernel.... [05:24] ah i see.. [05:24] cool.. [05:27] that makes sense [05:39] likewise, after you do a bunch of stuff, then I say 'baz switch kernel-team@...; baz merge zulcss@...' [05:40] yeah but i dont have sftp access [05:41] that's why I said 'I' :-) [05:42] heh [06:09] morning [06:24] did my 'Breezy development discussion' email to kernel-team make it out to everyone? [06:24] or was I the only one who got it when I sent it? [06:24] i got it... [06:30] good enough [06:31] dup elimination, followed by no traffic, can lead to worries... [06:32] yay [06:32] I got it, will respond [06:40] lamont: readding sparc32 is pretty straighforward [06:40] the point is that i don't have anyway to test the kernel at all [06:40] if you can install ubuntu with a debian kernel [06:40] i can readd the kernel and you can test it [06:41] or give me console access to the machine :) [06:41] and a netboot setup [06:44] yeah - it's not high on my list, atm [06:45] yeah i can understand that [07:20] dilinger: i am not going to pull in any of the 198 patches from andrew [07:20] they are all pushed to linus [07:20] and i am pretty sure linus will merge most of them blindly === atc [~stephen@69-161-205-34.clspco.adelphia.net] has joined #ubuntu-kernel === atc [~stephen@69-161-205-34.clspco.adelphia.net] has left #ubuntu-kernel ["Leaving"] === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-kernel === Xgates [~Xgates@cpe-70-93-51-212.hawaii.res.rr.com] has joined #ubuntu-kernel [09:29] hey all [09:29] anyone in? [09:29] I'm a Slacker compile junkie of 5 years, but was looking for some more ease of use and starting to move away from some of the compiling, but I still love it [09:29] I just compiled a patched source of 2.6.10, and got all the correct hardware compiled in just for my box only so its pretty stripped down like a vanilla kernel [09:29] everything boots up ok with the kernel, but at the login prompt the keyboard is dead and for the life of me for the past 3 days I cant figure why the keyboard is not working [09:29] it's just a PS/2 and I have it compiled in as --> [*] Keyboards AT keyboard support [09:29] what heck might there be that Ubuntu wants to get this thing going? [09:30] Xgates: if you compile your own packages, you need to get the config right [09:30] otherwise just use the normal packages [09:30] the config is right [09:30] thats the thing [09:30] [*] Keyboards AT keyboard support is al you need to run a PS/2 keyboard in Linux [09:30] Xgates: if your keyboard doesn't work, the config is not right :) [09:31] did you build the initrd? [09:31] note that it is a module [09:31] anything beyond that is not normal Linux as a Vanilla kernel goes [09:31] it needs to be loaded [09:31] I have built it with and without [09:31] sorry what as a mofule? [09:31] module? [09:31] initrd? [09:32] [*] Keyboards [09:32] either you load the module or you create the initrd [09:32] the ubuntu kernel as it is needs an initrd [09:32] ok right now I did not compile in ramdisk, initrd [09:32] otherwise change M to Y [09:32] I mean you should be able to run Ubuntu with out them [09:32] no [09:33] ubuntu can run without [09:33] but than you need to compile your kernel properly [09:33] no and what? [09:33] as i said your config is wrong [09:33] [*] Keyboards AT keyboard support works in a kernel without initrd [09:33] are you saying it does not in Linux? [09:33] Xgates: M -> module [09:33] because I have been doing it this way in Slackware for 5 years just fine [09:34] if you don't load the module [09:34] you don't get the keyboard [09:34] sorry what module? [09:34] the keyboard module [09:34] [*] Keyboards [09:34] OR [09:34] the one that makes the keyboard working [09:34] AT keyboard support ? [09:34] AT keyboard support [09:35] I did have it as a Module and compiled in <*> [09:35] and both ways the keyboard would not work [09:35] Xgates: does the keyboard work with the shipped kernel? [09:35] yes thats the thing [09:35] and I have compiled in all the main things that Ubuntu should need [09:35] Xgates: than there must be some other options that needs to be changed from M to * [09:35] I compiled them the same way [09:35] FS support, ramdisk, etc... [09:35] al same [09:36] all.... [09:37] Xgates: well.. there must be a mistake than... [09:37] thats the thing there is really nothing here that should effect the keyboard [09:37] it's obvious :) [09:37] yea sure but it's nothing to do with normal keyboard support in any other Linux distro [09:38] a kernel in Slackware is a Kernel in Debian as in Gentoo, etc... [09:38] Xgates: our kernel is no different in that respect [09:38] the only time something isnt going to work is because of something comfigure different into the distro, [09:38] configured.... [09:39] sorry I dont mean that Ubuntus kernel is different [09:39] I mean Ubuntu is [09:39] Xgates: the way Ubuntu kernel is managed is exactly the same as Debian [09:40] so if Debian works Ubuntu does and so on... [09:40] yes but not like Slackware [09:40] well we are different distros [09:40] there are differences [09:40] I have the exact same 100% settings in this kernel as it as been in Slackware for the past 3 years and now I cant get it to work in Ubuntu [09:40] we use an initrd boot system [09:40] i don't know about slack [09:40] yes thats what I mean [09:41] and running udev, etc.. which isnt the Slackware norm [09:41] these things make things like what Im encountering as the problem [09:41] learning the distro quirks as I call them [09:42] Xgates: if we were all using the same, there would be no reason for Slack or ubuntu or debian to exists [09:42] each distro is unique in its desig [09:42] design even [09:42] + i would really suggest you to use the standard kernel [09:42] there is really no reason to recompile it [09:43] fabbione: right now I dont have ram, initrd support in it --> < > RAM disk support [09:43] unless Ubuntu has a hard time without it [09:44] Xgates: again.. if you remove that options you need to make sure to compile the kernel properly. [09:44] yes I have [09:44] otherwise enable them and use initrd [09:44] been doing it 5 years just fine until I started using Ubuntu [09:44] hehe [09:45] fabbione: would you be so kind to look at this .config and see if you see anything besides my hardware setup and the basic kernel setups IF there might be something that Ubuntu is going to be looking for [09:46] Xgates: put it somewhere, but a) i won't have time to compile and build b) i dunno when [09:47] I dont need you to compile it [09:47] hehe [09:47] I just want you to look at it is all if you dont mind and see if you notice anything that Ubuntu needs that I missed [09:48] because I did not miss the correct settings for the kernel or my hardware [09:48] this is the same box I've been running for 3 years and compiling on it 3 years with the same kernel settings [09:48] one sec let me upload it [09:52] fabbione: http://www.patientswithouttime.com/2.6.10 [09:54] CONFIG_KEYBOARD_ATKBD=m [09:54] this needs to be y [09:55] fabbione: you mean --> AT keyboard support [09:56] if so I had it compiled in <*> before [09:56] and still didnt work [09:56] but then again I might of had ram and initrd with it too [09:56] I forget now [09:56] CONFIG_SERIO_PCIPS2=m [09:57] CONFIG_SERIO_I8042=m [09:57] all these stuff must be yes [09:57] config SERIO_I8042 [09:57] tristate "i8042 PC Keyboard controller" [09:57] wel I was just dorking with those BUt I dont need those I have a PS/2 [09:57] wel/well.... [09:58] config SERIO_RAW [09:58] tristate "Raw access to serio ports" [09:58] help [09:58] this one too [09:58] yes you do [09:58] read the help [10:00] ok my bad --> <*> i8042 PC Keyboard controller [10:00] ok well why does Ubuntu need these compiled in, and why wont it take them as ? [10:00] Xgates: as i explain to you before [10:01] the initrd stuff takes care of loading them when needed [10:01] so if you customize your config you need to change it properly [10:02] ok so in a initrd distro like Ubuntu, Debian, if you compile in ramdisk, initrd then Mods are ok BUT if you dont use them then you need to compile in <*> these for the support, keyboard, etc...? [10:02] then I need to do the same for the mouse yea --> PS/2 mouse [10:02] change to <*> [10:02] man Im not use to using a initrd system [10:02] Xgates: that is correct [10:02] hehe in Slack as a is just fine [10:02] ahh ic ok THANKS [10:03] sheesh [10:03] hehe :-) [10:03] I've run Debian but thats when I was starting out in Linux so for me the past 5 years have been in Slackware [10:04] I never got into initrd systems so I understand them somewhat but didnt know this about them with the kernel ;-) [10:05] but what are we saying NOTHING can be compiled as a ? [10:05] what about this --> Normal floppy disk support [10:05] boy this initrd thing is confusing me [10:05] hehe [10:05] I mean IF you want to use things as a module for anything then you have to use the ramdisk, initrd in Ubuntu [10:06] so in a nutshesll not ramdisk, initrd support then NO modules at all? [10:06] nutshell..... [10:06] errr not/no ram ...... [10:08] you can still use modules, but you need ensure that all the stuff you need to boot your machine is compiled in or loaded via /etc/modules [10:08] ok yeah Slacks the same in that respect [10:08] Xgates: that's the same everywhere :) [10:09] it's just how the kernel works [10:09] some distros use one approach [10:09] others the other :) [10:09] yes I know :-) [10:09] ahh /etc/modules is like SO different then in Slack [10:09] haha [10:09] sheesh just add in the name [10:09] hehe [10:10] so I guess just put in --> floppy [10:10] :_) [10:10] :-) [10:10] man I tell ya moving around from distro to distro can be confusing learning the ways in which the DEV made a distro, it's tools and quirks [10:10] hehe [10:12] i have been there a few years back [10:12] ;p [10:12] crap [10:12] VIA chipset support [10:12] it doesnt list the mod name [10:13] so I wouldnt know what name to put in /etc/modules to load it [10:13] crap [10:13] hmm [10:13] oh forgot it I'll make sure ram, initrd is compiled in [10:13] :-) [10:17] fabbione: hey one last one, you make a kernel with 'kernel package' and then install it, it boots and works, so then you remove the default kernel, NOW if you apt-get nvidia-glx will it install it into the /lib/modules path of this kernel you made, also I tried to install nvidia-glx, but then it was looking for the default kernel to install it back in, how do you get around all this? [10:17] apt-get install nvidia ...... [10:18] btw THANKS for the help :-) [10:18] Xgates: nvidia-glx is user land [10:18] you will have to recompile the kernel module [10:18] user land? [10:19] recompile the kernel module what ya mean? [10:19] Xgates: please ask these questions in #ubuntu [10:19] a lot of people knows how to do that [10:19] and i am kinda busy to explain these concepts [10:19] hehe well since I'm working on a Kernel thats why I asked there [10:20] nvidia is made of 2 parts [10:20] kernel and userland [10:20] no I have not been able to get any compile issues like this solved in #ubuntu [10:20] yes but this isn't really a support channel [10:20] this is where we discuss the kernel development [10:21] ok well if you can bare with me 5 more mins I'll be gone I really appreciate it [10:21] anyway i will answer your last question :) [10:21] ok [10:21] you need to install nvidia-glx to let X understand you have the binary crap installed [10:21] I have installed and compiled Nvidia on Slackware if this is what your suggesting [10:22] well than you do exactly the same on Ubuntu [10:22] ok wait backup [10:22] and don't install nvidia-glx if you use the installed from nvidia [10:22] not sure you missed what I said said [10:22] installer even [10:22] Xgates: if you use the nvidia installer from nvidia you don't need to install nvidia-glx package [10:22] as simple as this [10:23] I mean if I make my own kernel, then rip out the default and THEN want to install nvidia --> apt-get install nvidia-glx it wants to install the default kernel image back that I dont want [10:23] yes that's why i said don't install nvidia-glx [10:23] yea sure just thought I'd try to keep it all .deb Ubuntu style [10:23] Xgates: isntall nvidia-kernel-source [10:23] ahhh [10:23] and use module assistant to build it [10:23] or something like that [10:24] that will provide the bits you need iirc [10:24] i don't play with nvidia [10:24] ok [10:24] so i simply don't know [10:24] ok thanks [10:27] fabbione: THANKS alot for the help for this initrd run-down :-) [10:32] np [10:49] UHHAAAAA [10:49] * Add cluster support from Red Hat: [10:49] - Add Global Network Block Device (GNDB) support: [10:49] . Add patch external-drivers-block_gndb.dpatch and [10:49] external-drivers-block_gndb-fixups.dpatch. [10:49] . CONFIG_BLK_DEV_GNBD=m for all arches. [10:49] gotta love these CRACK! [12:50] Mithrandir: i am not sure the patch is right [12:50] it has a couple of extra fields compare to what pci_device_id can take [12:50] fabbione: it might be totally wrong, I'm not a kernel hacker. :) [12:50] oh don't worry.. i am just trying to understand :) [12:54] there is something weird [12:55] +i2o_block 0x00001044 0x0000a511 0x00001044 0x0000c034 0x00010400 0x0 0000000 0x0 [12:55] this entry defines 8 fields after the module name [12:55] but i have only 7 in the include [12:55] Mithrandir: can you paste me the first line of modules.pcimap please? [12:57] # pci module vendor device subvendor subdevice class clas [12:57] s_mask driver_data [12:57] hm [12:57] oh i seee.. [12:57] might be a paste error: [12:57] i2o_block 0x00001044 0x0000a511 0x00001044 0x0000c034 0x00010400 0x00000000 0x0 [12:57] is the line [12:57] ah [12:57] sorry, mea culpa [12:58] no problem :) [12:58] so it's probably a 0, too much in my patch [01:00] not only.. i need to understand the meaning of 0x00010400 and why 0x00001044 0x0000c034 in 3rd and 4th entry [01:00] PCI_ANY should give 0xffffffff [01:00] I think those can just be masked out [01:00] yes, but I changed that, since I don't think they should matter. [01:00] can you try changing the values and see if it works? [01:00] so we don't narrow it too much. [01:00] ah ok [01:01] sorry that wasn't clear. [01:06] actually.. [01:07] the fix is not that either [01:09] oh? [01:10] check drivers/message/i2o/pci.c [01:10] is the i2o_core module loaded properly? [01:11] it looks to me like the pci_table entry in pci.c is just wrong [01:12] i can't see any entry of it in pcimap [01:18] root@:~ # lsmod | grep i2o_core [01:18] i2o_core 46000 1 i2o_block [01:18] the pcimap was hand-written by jbailey [01:19] that is correct because i2o_block will pull in i2o_core [01:19] but it should happen the other way around [01:20] that i2o_core recognize the pci device [01:20] and load i2o_block [01:20] uhm, isn't core just a support thing? [01:20] i think i have a patch :) [01:20] Mithrandir: Ah, did I get it wrong? Sorry. =) [01:20] jbailey: not sure.. but it looks strage [01:20] jbailey: I don't think so, no, argue with Fabio :) [01:21] jbailey: I'm abusing your ia64 a little, since I managed to fuck up the previous upload :( [01:21] fabbione: I didn't try the case of jus tloading the control and hoping. =) [01:21] Mithrandir: That's fine. =) [01:21] I guess I'll have to buy bdale some beers. [01:21] Mithrandir: It may as well do something for all the electricity it consumes. [01:23] fabbione: I'm guessing if we let it auto load the drivers that we may eventually face the same problem we have with ide and scsi, which is that I need a hotplug pass to iterate over the bus. [01:24] fabbione: If there's an i2o device that needs a driver in the initrd it'll just silently not load. [01:24] jbailey: i think the problem is way different here [01:25] the pci_table_id in i2o/pci.c is simply broken [01:25] fabbione: Ah, but that still might be an issue down the road for coldplugging. [01:26] jbailey: if it is broken we fix it :) [01:26] fabbione: Yes, dear. =) [01:26] Mithrandir: http://people.ubuntu.com/~fabbione/drivers-message-i2o-i2oblock.dpatch [01:26] can you please compile a kernel with that patch? [01:27] and see if it solves the problem? [01:27] because neither PCI_DEVICE_CLASS or PCI_DEVICE fills up the table properly [01:28] sure, give me a few minutes [01:29] i have no rush [01:29] i am trying to get DLM to build [01:29] cluster/dlm/nodes.c:280: error: syntax error before '&' token [01:29] static void nodes_clear(struct list_head *head) [01:29] i can't see anything obviously wrong here [01:29] but gcc still errors on it [01:40] it must be one of these very well hidden typos [01:43] our kernel needs TeX to build. [01:43] oh well [01:44] Mithrandir: for the documentation... [01:44] still insane [01:44] :) [01:44] that should be changed to Build-Indep [01:45] isn't the kernel compile using -j ? [01:45] it seems a bit slow here [01:47] Mithrandir: CONCURRENCY_LEVEL [01:48] either you export it or uncomment it in debian/rules [01:48] a tad better, yes. [01:51] ehhe [01:51] btw.. how are you building? [01:51] dpkg-buildpackage? [01:53] yeah [01:53] after removing the variants I'm not building [01:53] ah ok [01:54] if you call ./debian/rules binary-debs is better [01:54] it will avoid trying to build the udebs [01:54] the debs will be in debian/build === fabbione kicks dlm in the middle of the code [02:00] bah this piece of code is driving me nuts [02:19] Uh oh. My laptop obviously knows we're going travelling soon... [02:19] hda: drive not ready for command [02:19] pauses the middle of my bootup. [02:19] *phear* [02:30] ARGHHHHH === fabbione hits everything [02:31] there was a name clashing in 2 functions [02:36] oh never mind [02:37] ops [02:37] fabbione: didn't work, it seems [02:37] crap [02:38] is there anything in pcimaps related to 0x1044 ? === zul [~chuck@198.62.158.205] has joined #ubuntu-kernel [02:46] hey [02:46] hey zul === zul checks the logs [02:46] i have to make sure you guys are not talking shit about me ;) [02:47] so we are actually using gcc-4 now for breezy? [02:48] zul: for userland they already switched [02:48] i am not sure what we will do for the kernel yet [02:49] and is the external drivers the latest one? [02:49] er..most uptodate [02:49] zul: if you have the cluster patches yes [02:50]