WebHosting Paid by #1Payday.Loans
[00:07] <ringo> moin [00:08] Action: ringo = AdHd = a gift [00:40] daja77 (n=daja77@dslb-088-072-036-211.pools.arcor-ip.net) joined #rocklinux. [00:51] daja77_ (n=daja77@dslb-088-072-037-159.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out) [01:58] ringo_ (n=info@ip51cf58c9.direct-adsl.nl) joined #rocklinux. [02:04] ringo (n=info@ip51cf58c9.direct-adsl.nl) left irc: Read error: 110 (Connection timed out) [02:26] ringo_ (n=info@ip51cf58c9.direct-adsl.nl) left irc: Read error: 110 (Connection timed out) [02:36] kasc_ (n=kasc@dslb-084-060-097-244.pools.arcor-ip.net) joined #rocklinux. [02:38] blindcod1r (i=blindcod@tor/session/direct/x-85c38402b1ca3c18) joined #rocklinux. [02:38] blindcoder (i=blindcod@tor/session/direct/x-946c61c0af4ebfa7) left irc: Nick collision from services. [02:39] Nick change: blindcod1r -> blindcoder [02:44] kasc (n=kasc@dslb-084-060-096-206.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out) [02:44] Nick change: kasc_ -> kasc [05:40] mnemoc (n=amery@216.241.24.10) left irc: Read error: 110 (Connection timed out) [06:15] mnemoc (n=amery@216.241.24.10) joined #rocklinux. [07:24] <blindcoder> moin [08:25] mnemoc (n=amery@216.241.24.10) left irc: Read error: 104 (Connection reset by peer) [08:25] mnemoc_ (n=amery@216.241.24.10) joined #rocklinux. [08:26] Nick change: mnemoc_ -> mnemoc [11:07] <esden> moin [11:40] <blindcoder> moin esden [12:55] <daja77> i am shocked, no new kernel release [13:00] <th> slackers [13:11] <blindcoder> heh [15:17] blindcoder (i=blindcod@tor/session/direct/x-85c38402b1ca3c18) left irc: Remote closed the connection [15:21] blindcoder (i=blindcod@tor/session/direct/x-0b78a507385e425b) joined #rocklinux. [15:49] <fake> moin [16:18] <fake> wooo bist duuuu meeein soooooooonnenliiiiicht [16:18] <fake> ich vermiiiiisseeeee diiiiisch [16:18] <fake> >_< [16:21] <th> fake: dont drink and irc. [16:22] <daja77> fake: i am here :> [16:22] <daja77> cnr [16:23] <fake> ich reeeeheeespeeeektier nur diiiiisch... [16:23] <fake> ich vermisse deinem aaaaaten [16:23] <fake> *rofl* [16:23] <daja77> hehe [16:24] <daja77> perhaps he got drunk from it [16:26] <fake> no [16:26] <fake> i'm 100% sober [16:28] <daja77> what a pity [16:31] <blindcoder> re [16:31] <blindcoder> th: ping [16:31] <th> blindcoder: pong [16:31] <blindcoder> th: before I'm going to snap: what is it you don't like about current rockinitrd and why do you think it needs changing? [16:32] <th> blindcoder: i just needed a initramfs to boot one of my systems. so i quick-created one... [16:33] <th> blindcoder: then i needed to change it a bit. and did some scripts [16:33] <th> and then i thought... hey another 15 minutes and this could be even more cool [16:33] <th> i did not plan to do it for rock but only as a quick-n-dirty solution [16:33] <th> but it looked pretty cool [16:33] <th> so i polished it further [16:33] <th> and i think the result is pretty slick [16:34] <th> i dont insist on changing the current way [16:34] <th> but i have this thing - and we could try it. [16:34] <blindcoder> th: I meant that question wrt. the line you posted yesterday [16:34] <th> for example loopback devices [16:34] <th> and it lacks modularity [16:35] <th> like the build.d/ directory in my approach [16:35] <th> the current approach is about filesystems [16:35] <blindcoder> modularity? [16:35] <blindcoder> ehm... [16:35] <th> we dont need any filesystem [16:36] <blindcoder> you know, I have this nice little way of creating initrd plugins... [16:36] <blindcoder> which don't need to do _anything_ with filesystems _at_ _all_ [16:36] <th> i took 10 minutes to inspect the current rockinitrd before i decided to just put a quick initramfs together with gen_init_cpio [16:36] <blindcoder> in fact I even have a plugin here to activate both LCD and CRT outputs on my laptop [16:36] <blindcoder> hardly a filesystem thing [16:37] <blindcoder> I agree that we should move to initramfs if the kernel is going to drop initrd anytime soon [16:38] <th> i dont think that it is. [16:38] <th> i'm talking about mkinitrd.sh [16:38] <th> initrdfs="cramfs" [16:38] <th> ramdisk_size=8192 [16:39] <th> that's just not necessary with initramfs - and i'm not sure if adding another option to rockinitrd to create cpio is the best option [16:39] <th> the mkinitramfs thing is so small - compared to this [16:39] <blindcoder> btw: https://www.rocklinux.org/wiki/Rock_InitRD_Concept [16:39] <blindcoder> th: well, the selection between filesystems is stf's thingie [16:40] <blindcoder> if it was me I'd just add squashfs to the kernel for _all_ distributions and use that [16:40] <th> the whole copying of /etc/conf/kernel is just not needed [16:40] <th> emit_udev_events() that's just obsolete [16:41] <blindcoder> then the size and fs-type are equally of no concern [16:41] <th> echo "loading kernel modules" [16:41] <th> . /etc/conf/kernel [16:41] <th> obsolete again [16:42] <th> i dont want to offend in any way... i just thought it might be interesting for some people to see what i did for mkinitramfs [16:46] <blindcoder> well, i haven't had a look at udev in a long time, it's just changing too fast for me to keep up [16:47] <blindcoder> when I created the code (which isn't the current code) it was the accurate way of handling things [16:50] <th> yea [16:51] <blindcoder> so [16:51] <blindcoder> udevd --daemon [16:51] <blindcoder> udevtrigger [16:51] <blindcoder> udevsettle [16:51] <blindcoder> does all udev stuff and module loading? [16:51] <th> yes [16:51] <blindcoder> hardware-autodetection? [16:51] <th> yes [16:51] <th> well half of the job does depmod [16:52] <th> by creating the modules.* files [16:52] <th> modules.alias [16:52] <th> maps pci-id's to module names [16:52] <th> so you can do "modprobe MY_PCI_ID" [16:52] <th> and you get the corresponding module loaded [16:52] <th> so the task for udev is easy [16:52] <blindcoder> okay, let me just finish encrypting my harddisk [16:54] <th> i wont hold you back [16:54] <blindcoder> yes, but do I have to write that modprobe stuff somewhere? [16:54] <blindcoder> will udev automatically load my USB modules when I plug something in? [16:54] <th> which stuf? [16:54] <th> i think so [16:54] <blindcoder> my nvidia module? centrino-wlan? PCI-Ethernet at least? [16:55] <th> if you nvidia module is listed in your /lib/modules/`uname -r`/modules.alias file. then yes [16:55] <th> s/you/your/ [16:57] roura-cz (n=root@80.95.103.208) joined #rocklinux. [16:59] <blindcoder> sweet [16:59] <blindcoder> okay, I'll try to put this stuff into the current initrd and remove the rest [17:00] <blindcoder> what about putting squashfs into the kernel by default? [17:04] <th> why? [17:04] <blindcoder> to get rid of fstype="..." and initrdsize="8192" [17:05] <th> well i've no squashfs in kernel and no fstype="..." [17:06] <blindcoder> 16:33 < th> initrdfs="cramfs" [17:06] <blindcoder> 16:33 < th> ramdisk_size=8192 [17:07] <th> yes? [17:07] <blindcoder> we could probably get rid of that using squashfs [17:07] <blindcoder> then again... maybe not [17:07] <th> we can get rid of it without squashfs [17:08] <th> we dont need ANY filesystem [17:09] <blindcoder> yes, for initramfs we don't [17:09] <th> exactly [17:10] <blindcoder> I would like to keep the current plugin-system, though [17:10] <th> i'm not pushing you to remove anything, am i? [17:11] <blindcoder> I'm not saying that :P [17:11] <th> good ;) [17:13] <blindcoder> btw, your mkinitramfs.sh still need root-access [17:16] <th> no [17:16] <th> only if you want to put it to a location where you are not allowed to write [17:17] <th> like if you dont change the default outfile location and your /boot is write protected [17:17] <blindcoder> running depmod? [17:17] <th> you need root for that? [17:17] <blindcoder> ./build-list.sh: line 10: depmod: command not found [17:17] <blindcoder> mkinitramfs.sh: line 79: /boot/initramfs-2.6.15.1-rock.cpio.gz: Permission denied [17:17] <th> it's only enumerating through a directory [17:17] <th> blindcoder: specify another outfile location [17:17] <th> or get yourself write access to your /boot [17:18] <blindcoder> still: ./build-list.sh: line 10: depmod: command not found [17:18] <th> mkinitramfs.sh --help [17:18] <th> mkinitramfs [ -r KERNEL_VERSION ] [ -m MODULES_DIR ] [ -o OUTPUT_FILE ] [17:18] <th> If no options are given the following defaults apply: [17:18] <th> mkinitramfs -r 2.6.12.6 -m /lib/modules/2.6.12.6 -o /boot/initramfs-2.6.12.6.cpio.gz [17:18] <th> bash ./mkinitramfs.sh -o /tmp/foo [17:18] <th> ... [17:18] <blindcoder> /sbin/depmod [17:18] <blindcoder> that's not in a normal users path [17:18] <th> what's the problem with depmod? [17:19] <blindcoder> it's in /sbin/ [17:19] <th> ok - so i'll make /sbin part of PATH in mkinitramfs [17:20] <th> export PATH=$PATH:/sbin:/bin:/usr/bin:/usr/sbin [17:20] <th> thanks for reporting. it's fixed [17:21] <th> i propably did not notice because i have it in my path [17:21] <th> blindcoder: please try to run "bash ./mkinitramfs.sh -o /tmp/foo" and tell me that you DONT need root privs [17:24] <blindcoder> th: I don't need for that [17:24] <th> good [17:25] <th> so you propably see that your "< blindcoder> btw, your mkinitramfs.sh still need root-access" does not hold true, correct? [17:26] <blindcoder> it almost does :) [17:26] <th> nope. [17:26] <blindcoder> anyway, AFAICS gen_init_cpio only creates a normal CPIO file [17:26] <th> what's the problem with a normal cpio file? [17:27] <blindcoder> so we can basically do the same with 'cpio -o < <( find ) > /boot/initramfs' [17:27] <th> i doubt that. [17:27] <th> try to add a device node like that [17:27] <blindcoder> root@fuzzy:~/tmp# ll dev/ [17:27] <blindcoder> total 0 [17:27] <blindcoder> crw------- 1 root root 5, 1 Apr 21 17:15 console [17:27] <blindcoder> root@fuzzy:~/tmp# cpio -t < /boot/foo | grep console [17:27] <blindcoder> dev/console [17:28] <blindcoder> root@fuzzy:~/tmp/tmp# cpio --no-absolute-filenames -i < /boot/foo [17:28] <blindcoder> 58655 blocks [17:28] <blindcoder> root@fuzzy:~/tmp/tmp# ll dev/ [17:28] <blindcoder> total 0 [17:28] <blindcoder> crw------- 1 root root 5, 1 Apr 21 17:22 console [17:28] <blindcoder> /boot/foo was created like this [17:29] <blindcoder> root@fuzzy:~/tmp# cpio -c -o < <( find ) > /boot/foo [17:29] <th> cpio: .: truncating inode number [17:29] <th> cpio: null: truncating inode number [17:30] <blindcoder> root@fuzzy:~/tmp# mine -q cpio [17:30] <blindcoder> cpio 2.6 0 [17:30] <th> i see quite some advantages in gen_init_cpio: [17:30] <th> it's promised to be compatible with linux kernel [17:30] <th> we dont need root privs [17:30] <th> we dont need cpio [17:31] <th> lemme emerge a more recent cpio [17:31] <blindcoder> I just wonder how to translate the current /etc/conf/initrd/initrd_* files [17:32] <th> bye doing sed 's/\(.*\)/echo file \1 0755 0 0/' ? [17:33] <th> and putting them in build.d [17:34] <th> well [17:35] <th> might be easier to add them like done in build.d/05-bins [17:35] <blindcoder> I'd favor them like this: [17:35] <blindcoder> while read file ; do stat <bla> $file ; echo <correct output> ; done < /etc/conf/initrd/initrd_* [17:36] <th> have you seen 05-bins? [17:36] <blindcoder> just looking at it [17:36] <th> perhaps there is a >>> test -z "`which $x`" <<< missing [17:36] <blindcoder> 0755 might not always be correct [17:37] <blindcoder> it will work, but I'd prefer my config files not to be executable :) [17:37] <th> well 05-bins is for binaries. [17:37] <th> 06-conf or something could have 0644 or something [17:37] <blindcoder> yeah, but /etc/conf/initrd/initrd_* doesn't contain just executables [17:37] <blindcoder> and I'm quite fond of the format those files have [17:38] <th> the gen_init_cpio format is almost the same [17:38] <blindcoder> anyway, I think I have a good view on how to create the initramfs [17:38] <th> but it contains filetype, permission, and ownership [17:38] <blindcoder> I'll merge the rockinitrd with your mkinitramfs and see what happens :) [17:39] <blindcoder> in turn I'd like you to apply 2006031713044803603 later [17:41] <th> in turn?? [17:41] <blindcoder> :) [17:41] <th> i'm not in your death or something? [17:41] <blindcoder> I sure hope not :) [17:41] <th> cool then ;) [17:42] <th> well we could move rockinitrd back to blindcoder and add mkinitramfs to base ;-) haha [17:42] <th> then i'll apply just anything - cause it's public/personal :-) [17:43] <th> oh that is recreating initrd from inside it [17:44] <blindcoder> or we could just dump rockinitrd when rockinitramfs is done [17:47] <daja77> omg sometimes i hate myself [17:47] <blindcoder> rest assured, so do we :) [17:47] <th> hmm? [17:47] <th> how rude [17:48] <blindcoder> stop taking everything so serious today [17:48] <daja77> i just created a nice image for diploma, and on they way i destroyed all the data i needed to create it, or well i lost them [17:48] <daja77> like i have been on drugs [17:48] <th> blindcoder: you just needed some serious treating ;) [17:49] <daja77> just i don#t remember any drugs, apart from coffeine [17:51] Action: blindcoder goes to watch his lava lamp then [17:53] <blindcoder> th: what is /sbin/vol_id? [17:53] <blindcoder> th: it seems to be missing from my current initrd [17:54] <th> blindcoder: that's a udev helper [17:54] <th> blindcoder: but it should be in /lib/udev [17:54] <daja77> sounds like an afs program [17:54] <daja77> <- off [17:54] <th> cya daja77 [17:55] <blindcoder> udevd-event[1675]: run_program: exec of program `/sbin/vol_id` failed [17:55] <blindcoder> will look at that later then [17:55] <th> blindcoder: have you applied 2006041801092016294? [17:55] <th> see https://www.rocklinux.net/people/teha/mkinitramfs.txt [17:55] <blindcoder> bye daja77 [17:56] <blindcoder> hmm, I don't think so [17:57] <th> see https://www.rocklinux.net/people/teha/mkinitramfs.txt then [17:57] <th> and fix your udev by doing this. [17:57] <blindcoder> will do [18:28] roura-cz (n=root@80.95.103.208) left irc: Remote closed the connection [18:38] roura-cz (n=grml@80.95.103.208) joined #rocklinux. [19:47] roura-cz (n=grml@80.95.103.208) left #rocklinux. [19:50] <blindcoder> *ARGH* [19:50] <blindcoder> cpio extracted its stuff into / !!!!!!! [19:56] <icelbox> hehe [19:56] <blindcoder> that's not funny [19:57] <blindcoder> my fstab is gone >_< [20:00] <icelbox> save early, save often... [20:31] <blindcoder> the fstab was done ages ago [20:34] <icelbox> hmmm, good that I am doing a full backup every 1st of the month ^^ [20:49] <blindcoder> not on this machine, i'm afraid :( [00:00] --- Sat Apr 22 2006