=== bjf is now known as bjf-afk [03:35] hi all, I need to install sudo apt-get install linux-restricted-modules-`uname -r` [03:35] and I dont understand why it needs to install a nvidia kernel [03:35] no way to have madwifi without the nvidia stuff ? [03:41] Where can I find nightly builds of the kernel? [03:45] http://kernel.org/ [03:48] so you don't have a snapshot build like the fedora folks with git snapshots? [05:05] spstarr: http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-intrepid.git;a=summary [05:07] looks 'old' [05:07] .27? [05:07] in intrepid [05:07] * spstarr looks for karmic kernel [05:08] https://wiki.ubuntu.com/KernelTeam/FAQ [05:08] ubuntu is debian based [05:08] yeah i know [05:08] if you want cutting edge kernel and apps . . . I d recommend something like mandriva [05:08] but its less stable most of the time [05:08] well once .32 comes out i wont need cutting kernel [05:08] its only the r6xx stuff i want to use [05:08] you can also build your own kernel its not _so_ hard [05:09] sure, and have, taking ubuntu's .config [05:09] .32 will be out sooner than later [06:07] Hello all [06:08] I'm testing ubuntu karmic and wanted to ttry fs-cache [06:08] the read speedup for nfs is the thing i'm interested in because of a terminalserver configuration [06:09] i was looking at the kernel config and saw that the config option for CONFIG_NFS_FSCACHE is not set [06:10] is there a specific reason why it hasn't been enabled? [06:17] latest 2.6.31 has no fbcon [06:18] oh, sorry, no, it's just not loaded automagically anymore [06:22] which kills video signal if used with radeon kms [06:23] uh, that wasn't so clear, if radeon module is loaded with kms and no fbcon is loaded you get no signal for virtual terminals [06:26] hi all, for madwifi I need to install sudo apt-get install linux-restricted-modules-`uname -r` [06:27] and I dont understand why it needs to install a nvidia kernel, no way to have madwifi without the nvidia stuff ? [06:27] furtermore trying to install it fails : [06:27] linux-restricted-modules is a metapackage probably [06:29] i couldnt find another way to get madwifi [07:19] is itpossible to find a kernel with the ath_pci module on ubuntu [07:20] seems it ned a more recent kernel than those in ubuntu [07:46] ne0futur: have you tried the karmic kernel? === cjwatson_ is now known as cjwatson [08:43] amitk_: i need to upgrade all my distro to get the recent kernel ? [08:45] ne0futur: You're looking in the wrong direction, ath_pci is old and removed [08:46] ne0futur: no, not upgrade the distro. But you can get just the karmic kernel and see if that helps. It has the new ath5k drivers. [08:50] karmic kernel won't give madwifi, though it will give an up-to-date ath5k which might be a suitable replacement [08:50] Though that seems to be an inadvisable course of action without first trying linux-backports-modules [09:01] hum ill try to see how i can try this kernel [09:01] apt-cache search karmic gives nothing [09:03] ne0futur: what release are you running currently? [09:03] maxb makes a good point. Try linux-backports-modules for that release and see if that solves your problem. [09:05] ne0futur: If not, you can simply download the karmic kernel .deb from http://packages.ubuntu.com/karmic/linux-image-2.6.31-10-generic (depending on whether you run 32-bit or 64-bit) and try that. [09:08] /etc/debian_version says lenny/sid [09:08] i cant find a /etc/ubuntu-release [09:09] but i instsalled a ubuntu ;) [09:09] I dont want to upgrade distro cause i dont want kde4, kde3 is perfect for me [09:10] ah I found /etc/lsb-release [09:10] ne0futur: what does /proc/version_signature and /etc/lsb_release say? [09:10] says DISTRIB_ID=Ubuntu DISTRIB_RELEASE=8.04 DISTRIB_CODENAME=hardy [09:11] 2.6.24-24.60-generic [09:11] ne0futur: in that case best try the linux-backports-modules packages [09:11] ( downloading the .deb from http://packages.ubuntu.com/karmic/linux-image-2.6.31-10-generic is not ok with this ubuntu release ?) [09:12] ne0futur: that release is very old. I don't think anybody has tested such a new kernel with Hardy. [09:12] (another question is there a best kernel for a thinkpad laptop than the -generic kernel ? ) [09:13] hum ok [09:13] if I upgrade my ubuntu will I be forced to migrate to kde4 or is it possible to stay on kde3 ? [09:14] ne0futur: that is a question for #ubuntu. I think the default might be KDE4 but it might be possible to get KDE3 back. [09:14] ok thanks [09:15] i ll try the .deb first and then think on upgrading or using backports [09:15] ne0futur: I'd try backports first, the .deb next and then think of upgrading :) [09:15] my other problem is that I nmeed an old xorg to have my intel graphics card working well with flash/3d [09:16] I switched fron mandriva to ubuntu for this reason : [09:16] https://qa.mandriva.com/show_bug.cgi?id=53468 [09:16] apw: I've got a new branch, fsl-imx51-next, in my local tree. Should I overwrite the fsl-imx51 in the karmic git tree or just create a new temp branch for you to review? [09:16] qa.mandriva.com bug 53468 in Hardware "very slow flash and graphics when using intel 945GM video card" [Major,New] [09:17] ne0futur: you can try out new Ubuntu releases using the livecds/liveusb. So you can check it out without reinstalling your OS. [09:17] yup i ll try this [09:17] apw: I want to get this uploaded today if possible [09:17] thanks [09:21] amitk_, hiya [09:21] let me just check the current branch [09:23] amitk_, we seem to have one commit since the last tag, so assuming you have that, then you can push to the main branch [09:23] * amitk_ checks [09:26] apw: yes, I got the one from Tim removing ABI/version from vmlinuz name [09:28] then i can't see any issue pushing it it, is it a rebase? [09:28] i assume so, as the old one is -6.25 based [09:29] apw: rebased onto 878e32db158b972b443c5cd8370b2d52068bcaaf (master). Though I see 22 commits since then. [09:30] probabally rebase pretty simply onto the latest tag i'd assume as there aren't any rebases since then [09:30] but if you've tested that one, that seems like a major improvement [09:30] over -rc6 [09:30] apw: I don't see the -10 tags [09:31] sometimes i find you need to git fetch --tags origin to get them all [09:31] i don't know why sometimes it doesn't get them [09:32] git fetch -t still doesn't get me a Ubuntu-2.6.31-10* [09:32] http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-karmic.git;a=summary [09:32] nevermind, I am blind [09:32] yet they really are there, gitkweb shows them [09:32] I'll rebase onto -10.34 [09:35] these arm kernels really are a lot of work [09:36] apw: you should see the tree I'm about to push :) [09:36] am i going to barf? [09:39] heheh ... [09:41] amitk_, hopefully the rebase will be unremarkable to that tag [09:44] yeah, unremarkable. No config changes too. I'll just push the changelog, skip abi and module checks and push the branch [09:46] apw: I'm a bit confused. Should debian/abi be touched or debian.fsl-imx51/abi ? [09:47] abis are in the per branch space, so .fsl-imx51 [09:47] getabis is still broken [09:49] apw: then why are debian/changelog debian/control, etc. created? [09:49] those have to be in those places as they are used by build system [09:50] built stuff is mostly still in debian as a result [09:58] apw: test building once before pushing [10:17] cool [10:42] amitk, do you know if the imx51 package uses the very same postinst -generic uses ? [10:42] i dont get any link in / which makes alternate and netboot installs fail miserably since flash-kernel urgently needs these links [10:43] apw, ^^^ ? [10:44] http://paste.ubuntu.com/273349/ ... [10:44] ogra, as far as i know its a direct copy [10:44] ~ # grep symlink target/etc/kernel-img.conf [10:44] do_symlinks = yes [10:45] but according to the log (and my inspections) the symlink isnt created [10:45] ogra, do they get made on on the dove images? [10:45] no idea [10:46] NCommander, did you try out alternate installs on dove yet ? [10:46] ogra, isn't the install of the kernel failing anyhow [10:47] update-initrmafs is failing because it calls readlink [10:47] looking for /vmlinz and /initrd.img [10:47] apw: ogra: I see a slight diff [10:47] is that the three new postrm's ? [10:48] http://pastebin.ubuntu.com/273351/ [10:49] amitk, thats only error messages if i reasd it right [10:49] yeah, seems to be just messages [10:50] Sep 17 17:59:47 in-target: Setting up linux-image-2.6.31-100-imx51 (2.6.31-100.7) ... [10:50] after that the symlinks should exist [10:50] Sep 17 17:59:48 in-target: Failed to create initrd image. [10:50] Sep 17 17:59:48 in-target: dpkg: error processing linux-image-2.6.31-100-imx51 (--configure): [10:50] Sep 17 17:59:48 in-target: subprocess installed post-installation script returned error exit status 2 [10:51] Sep 17 17:59:49 in-target: Package linux-image-2.6.31-100-imx51 is not configured yet. [10:51] apw, yup, thats the fallout of the missing links [10:51] it seems to have failed in such a way that its not configured [10:51] Sep 17 17:59:48 in-target: readlink: missing operand [10:51] thats the actual error [10:51] update-initramfs looks for /vmlinuz [10:52] which the postinst of linux-image-2.6.31-100-imx51 should have created before running update-initramfs [10:55] ogra: and it works in the desktop install images? [10:55] yes [10:56] but desktop doesnt need flash-kernel ... even if it would have gone through update-initramfs, it would have failed in flash-kernel, the links in / are essential on armel [10:57] ogra, are tehy on the desktop images? [10:58] that log there, how was it generated, is there a command line to do that or something? [10:59] thats the output of d-i [10:59] the desktop images dont use d-i [11:00] but does the kernel installed there have it [11:00] also when you install kernels manually do they normally maintain the links [11:01] obviously they do in the main kernel, and the script is identicle [11:01] yes, the desktop images have /vmlinuz and /initrd.img [11:01] else they wouldnt be installable :) [11:02] so generally the install is making them then [11:02] (flash-kernel, as i said above ) [11:02] right [11:02] the linux-image postinst reads /etc/kernel-img.conf and creates them [11:03] in d-i installs /etc/kernel-img.conf is created by base-installer right before linux-image* is installed [11:03] in live setups /etc/kernel-img.conf is created by livecd-rootfs right before the linux-image-* package gets installed in the chroot [11:03] same procedure two different environments and outcomes [11:04] have you looked at an x86 alternate installer image? [11:04] perhaps they are all broken [11:05] heh [11:05] someone would have noticed that ... the log is from A6 testing :) [11:05] i produced it yesterday with the A6 candidate image [11:05] not necessarily perhaps its not important on the alternate if you don't have flash-kernel [11:06] given that i386 all passed i doubt they have that issue [11:06] flash-kernel is run a lot later [11:06] fair [11:06] though i'm not sure why the readlink error shows up [11:06] if [ -L /initrd.img -a -e /initrd.img ]; then [11:06] linktarget="$(basename "$(readlink /initrd.img)")" [11:06] fi [11:06] if [ -L /boot/initrd.img -a -e /boot/initrd.img ]; then [11:06] linktarget="$(basename "$(readlink /boot/initrd.img)")" [11:06] fi [11:07] neither of the files exist yet, so the if statement isnt fulfilled [11:07] (thats from update-initramfs) [11:07] so it shouldnt call readlink at all [11:08] so i'd say it can't be that readlink which failed [11:08] there is no other readlink involved i think [11:09] Sep 17 17:59:48 in-target: update-initramfs: Generating /boot/initrd.img-2.6.31-100-imx51 [11:09] Sep 17 17:59:48 in-target: readlink: missing operand [11:09] Sep 17 17:59:48 in-target: Try `readlink --help' for more information. [11:09] Sep 17 17:59:48 in-target: mkinitramfs: missing root /sys entry [11:09] Sep 17 17:59:48 in-target: mkinitramfs: workaround is MODULES=most [11:09] ogra@osiris:~/Devel/packages/base-installer-1.101ubuntu4$ grep readlink /usr/sbin/update-initramfs [11:09] linktarget="$(basename "$(readlink /initrd.img)")" [11:09] linktarget="$(basename "$(readlink /boot/initrd.img)")" [11:11] ogra, but thats not all of it [11:11] all of what ? [11:11] Sep 17 17:59:48 in-target: readlink: missing operand [11:11] Sep 17 17:59:48 in-target: Try `readlink --help' for more information. [11:11] Sep 17 17:59:48 in-target: mkinitramfs: missing root /sys entry [11:11] Sep 17 17:59:48 in-target: mkinitramfs: workaround is MODULES=most [11:11] Sep 17 17:59:48 in-target: mkinitramfs: Error please report the bug [11:11] Sep 17 17:59:48 in-target: readlink: missing operand [11:11] thats the error [11:11] that comes from hook-functions in initramfs [11:12] yes but there are more readlinnks in that file [11:12] which is run by innitramfs [11:12] mkinitramfs [11:12] outfile="$(readlink -f "$outfile")" [11:12] oh you are right [11:13] thats like /sys is broke [11:14] i see the error msg, buit i dont get why sys would be involved at all [11:14] there are errors about things in sys not being there [11:14] which are surly not right [11:14] and led to it exiting 1 [11:15] which led to the kernel install failing [11:15] if [ -z "${block}" ] || [ ! -e /sys/block/${block} ]; then [11:15] echo "mkinitramfs: missing ${block} root ${root} /sys entry" [11:15] echo "mkinitramfs: workaround is MODULES=most" [11:15] echo "mkinitramfs: Error please report the bug" [11:15] exit 1 [11:15] that was what it didn't cope with, not syaing the readlink is irrelevant, but its not what triggered the failure of kernel install directly [11:16] apw, yeah, just found that too [11:16] so i think that means the kernel links likely would have gotten made 'shortly' but initramfs blew chunks so it stopped [11:16] yup [11:18] so its an installer issue, sorry for bothering you guys ... i'll try to find out why base-installer didnt mount /sys [11:18] ogra, not sure if it was mounted or not hard to tell from the error [11:18] there may yet be an issue that /sys/block was missing for some kerenl reason, but see if you can find out if it was mounted [11:19] well, i see that it isnt mounted atm, [11:19] ahh [11:20] ~ # mount [11:20] rootfs on / type rootfs (rw) [11:20] none on /proc type proc (rw,relatime) [11:20] none on /sys type sysfs (rw,relatime) [11:20] tmpfs on /dev type tmpfs (rw,relatime) [11:20] devpts on /dev/pts type devpts (rw,relatime,gid=5,mode=620,ptmxmode=000) [11:20] /dev/sda1 on /target type ext4 (rw,relatime,errors=remount-ro,barrier=1,data=ordered) [11:21] /dev/sda1 on /dev/.static/dev type ext4 (rw,relatime,errors=remount-ro,barrier=1,data=ordered) [11:21] tmpfs on /target/dev type tmpfs (rw,relatime) [11:21] neither /proc nor /sys are mounted under /target atm [11:21] if thats 'mid install' then that seems broke [11:22] amitk_, is this an abi-bumper, i assume it is with this huge pile of patches [11:22] apw: yes, and adds a module or two. So you could ignore abi and module checks [11:23] ok [11:23] i'll do that [11:23] thanks [11:23] * amitk_ steps out for a while === csurbhi is now known as csurbhi-outFor40 [13:20] amitk, how are the chances of seein the imx51 upload today ? (/me si assembling the release team report and would love to strike the rtc issue from the buglist) [13:45] ogra, i have the tree and it should be uploading imminently [13:46] * ogra hugs apw [13:46] i've build tested it, and i assume amitk has boot tested the basic tree [13:47] its an abi bumper so it'll be a while churning through the works [13:47] ogra, do you have the agenda for todays meeting yet? [13:49] yep, lool forwarded it to me, i think robbie doesnt know about the replacement people [13:50] apw: bounced [13:50] lool, thanks === csurbhi-outFor40 is now known as csurbhi === bjf-afk is now known as bjf [21:19] Hi [21:20] Why do the mainline kernel builds have only source and header packages lately? [21:20] http://kernel.ubuntu.com/~kernel-ppa/mainline/daily/current/