[00:09] <fake> glibc 2.4 is out
[00:09] <fake> since 6th of march
[00:09] <th> yea. that's so out.
[00:09] <th> it's so 80th
[00:10] <th> oh
[00:10] <th> you mean it's released?
[00:10] <th> yea - fresh stuff.
[00:10] <th> we should go for it ;-0)
[00:10] Action: fake just began testing gcc 4.0.3
[00:11] <fake> * The LinuxThreads add-on, providing pthreads on Linux 2.4 kernels, is no
[00:11] <fake>   longer supported.  The new NPTL implementation requires Linux 2.6 kernels.
[00:11] <fake>   For a libc and libpthread that works well on Linux 2.4 kernels, we
[00:11] <fake>   recommend using the stable 2.3 branch.
[00:11] <fake> hooray!
[00:12] <stf^rocklinux> fake: who says that?
[00:13] <stf^rocklinux> ah, glibc
[00:13] <stf^rocklinux> fake: so we should keep glibc 2.3 for e.g. cross-builds...
[00:14] <stf^rocklinux> because some cross-builds seem to work only with libpthreads for now
[00:15] <stf^rocklinux> th: I suggest to rewrite the glibc in C++ ^^
[00:15] <th> stf^rocklinux: will do that on weekend.
[00:16] ringo (n=info@ip51cf58c9.direct-adsl.nl) joined #rocklinux.
[00:19] <stf^rocklinux> th: possible desktop files fix is here: https://www.rocklinux.net/submaster/smadm.cgi?i=2006032500183625143
[00:20] <th> -var_append confopt " " "--with-xinerama --enable-dnotify"
[00:20] <th> +var_append confopt " " "--with-xinerama"
[00:20] <th> hm?
[00:20] <stf^rocklinux> th: that is not really related, but kind of hard to patch separately
[00:20] <th> ok
[00:20] <th> what is dnotify?
[00:21] <stf^rocklinux> linux kernel directory notifications
[00:21] <th> ah ok
[00:27] <th> HA! found some public patches to apply
[00:30] <fake> th: i was just skimming through them
[00:30] <th> fake: finally you underestimated my today's pace!
[00:33] <th> fake: can you reproduce the liblow (from gpm) message in w3m?
[00:33] <th> (needs to be run in console or konsole i think)
[00:36] <th> how's the linux-2.6.16 + linux-libc-headers + xorg status?
[00:36] <th> should it all work?
[00:36] <th> when put together?
[00:37] <th> fake: stf^rocklinux ?
[00:37] <stf^rocklinux> th: the generic build shows no special errors so far
[00:37] <th> where is it?
[00:37] <th> stage[0] linux-headers?
[00:38] <stf^rocklinux> th: stage 5, shortly before kde is built
[00:38] <th> sounds cool
[00:38] <th> i'll try it in my next crystal journal
[00:40] <fake> th: yes, it shouldn't cause too much regressions
[00:40] <fake> *many
[00:41] <th> stf^rocklinux: you were tinkering with gpm lately... does this "*** err [lib/liblow.c(265)] \n strncmp/isdigit/option.consolename failed" error message sound familiar to you?
[00:43] <stf^rocklinux> th: no, but did you google for "gpm liblow"?
[00:45] <th> 3 rock hits on first page
[00:46] <stf^rocklinux> th: hannes' patch modifies liblow.c... (package/base/gpm/liblow-silent.patch)
[00:47] <th> oh
[00:47] <th> that's only making it a bit more silent ;-)
[00:47] <th> funny tho
[00:47] ringo (n=info@ip51cf58c9.direct-adsl.nl) left irc: 
[00:49] <stf^rocklinux> th: maybe that helps: http://crux.nu/portdb/?command=viewfile&repo=contrib&port=gpm&file=gpm-1.20.1-no-dumb-error-messages.patch ;)
[00:50] <th> stf^rocklinux: 2006032320051117797  does this fix a bug when building dbus/hal?
[00:50] <th> stf^rocklinux: so should i see it as a dependency?
[00:51] <th> /* do we really need this check ? */
[00:51] <th> hehe
[00:51] <th> i bet we don't with udev/devfs 
[00:53] <stf^rocklinux> th: that patch is not hal related
[00:53] <stf^rocklinux> th: but pciutils must be built before hal
[00:55] <fake> oh great. why is it that machines never come up once you closed the case and doctored them into a rack?
[00:58] <th> stf^rocklinux: oh so i better change that prio.
[00:59] <stf^rocklinux> th: 2006032500560327842
[01:00] <th> stf^rocklinux: oh thanks
[01:00] <th> -smap 2006032422504226659
[01:00] <th> +smap 2006032500560327842
[01:00] <th> ;-)
[01:04] <fake> FUCK
[01:04] <th> fake: lemme guess: hardware sucks?
[01:05] <fake> indeed
[01:05] <fake> the voice over ip machine didn't come up anymore after installing a second ISDN card
[01:05] <fake> kernel oops on loading the fritz card drivers
[01:05] <th> liblow-problem: that check fails because /dev/pts/10 != /dev/vc/0
[01:06] <fake> now i have to re-set the CMOS, because some bios option i toggled decided to wreck the complete vga output
[01:14] <fake> ... which is hard, since the 'clear cmos' jumpers are removed on this 'server board'
[01:15] kasc_ (n=kasc@dslb-084-060-107-194.pools.arcor-ip.net) joined #rocklinux.
[01:20] <th> stf^rocklinux: 2006032500183625143 (kde-libs) thing is broken
[01:20] <th> stf^rocklinux: two hunks for parse-config
[01:22] kasc (n=kasc@dslb-084-060-107-230.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out)
[01:22] Nick change: kasc_ -> kasc
[01:22] <stf^rocklinux> th: thanks, I'll remove one of them
[01:22] <th> i can fix it
[01:22] <th> it's trivial
[01:22] <stf^rocklinux> thanks
[01:25] <fake> hm. just ripping out the battery works just as good ;)
[01:29] <th> haha
[01:48] <th> new build started
[01:48] <th> gcc4
[01:49] <th> http://www.rocklinux.net/people/teha/2006-03-25-journal.txt
[01:49] <th> http://www.rocklinux.net/people/teha/2006-03-25-journal.diff
[01:49] <th> based on r7339
[01:49] <th> boot, bbs, crystal
[03:05] <fake> crap
[03:06] <fake> big, stinking, evil, fat pile of CRAP
[03:23] <stf^rocklinux> ?
[03:24] <fake> ISDN support on Linux
[03:24] <fake> mISDN is unifinished and as such buggy as hell
[03:24] <fake> vISDN isn't even near to be usable
[03:25] <fake> zaphfc/bristuff is an ugly, ugly hack
[03:25] <fake> and finally, the fcpci commercial drivers from AVM are money-generators
[03:25] <fake> (they only support one card / machine, so you have to buy an active card to get more than one port)
[03:26] <fake> and this is not even *considering* running on an SMP machine
[03:59] <fake> == 03/25/06 04:03:16 =[0]=> Finished building package glibc24.
[04:50] <fake> == 03/25/06 04:54:06 =[1]=> Finished building package glibc24.
[04:51] Action: fake leaving, cya
[05:21] SteffenP (i=steffen@p549949C9.dip.t-dialin.net) left irc: Read error: 110 (Connection timed out)
[05:22] SteffenP (i=steffen@p549977AB.dip.t-dialin.net) joined #rocklinux.
[05:25] madtux (i=miguel@pf0.hostarica.com) left irc: "Leaving"
[09:23] <blindcoder> moin
[09:57] <stf^rocklinux> moin blindcoder
[09:58] <blindcoder> moin stf^rocklinux 
[10:04] blindcod1r (i=id@tor/session/x-84df38349950fd78) joined #rocklinux.
[10:05] blindcoder (n=blindcod@tor/session/x-44b528894b8bd1f2) left irc: Nick collision from services.
[10:05] Nick change: blindcod1r -> blindcoder
[10:24] <owl> moin
[10:42] SteffenP (i=steffen@p549977AB.dip.t-dialin.net) left irc: Remote closed the connection
[11:37] <netrunner> fake: mISDN works fine
[11:37] <netrunner> s/fine/works/ :)
[11:52] <th> "mISDN works works"?
[11:52] <th> 03:27:11 < fake> mISDN is unifinished and as such buggy as hell
[11:52] <th> i have it running in production
[11:53] <th> oh i even contributed.
[11:54] <th> oh well that is only true for chan_misdn
[12:21] <th> 151 builds total, 151 completed fine, 0 with errors. (boot)
[12:21] <th> 208 builds total, 208 completed fine, 0 with errors. (bbs)
[12:21] <th> 338 builds total, 87 completed fine, 0 with errors. (crystal)
[12:45] <blindcoder> hmm
[12:45] <blindcoder> do I need to rebuild mplayer/xine after installing libdvdcss?
[12:47] <netrunner> blindcoder: reading mplayer.log it looks as if it would use its own copy of libdvdcss
[12:47] <netrunner> if this is only a placeholder then maybe not.
[12:48] <blindcoder> ah
[13:08] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) joined #rocklinux.
[13:48] <fake> netrunner: no, it by far does not work 'fine'
[13:49] <fake> th: how exactly do you use mISDN? with what hardware?
[13:51] <th> beronet hardware
[13:51] <th> hfcpci
[13:51] <th> 8x card pci
[13:51] <th> and at home a 1x hfcpci
[14:13] <th> can qemu use a block device as disk?
[14:13] <th> instead of disk image?
[14:14] <fake> th: i tried using a billionport hfcpci for voice mainly, misdn was very unstable, eventually crashing the box on use
[14:15] <fake> th: so i switched to avm cards, their binary drivers at least work reliably when you have only 1 card/box
[14:15] <th>  14:17:50 up 96 days,  1:39,  1 user,  load average: 0.01, 0.02, 0.00
[14:15] <th> no crashes since install
[14:15] <th> i guess 96 days ago i upgraded misdn.
[14:15] <fake> th: load on the card? call volume? PPTP in TE mode?
[14:16] <th> fake: 20people in the office
[14:16] <th> 3x s0 together as ptp
[14:16] <th> 1x s0 first as ptmp now ptp as well
[14:22] <th> oh we have lvm2 in CORE but not lvm-wrapper?
[14:25] <fake> th: maybe my board or the card itself was broken, but the whole mISDN/chan_misdn setup was a total mess
[14:25] <fake> may even be related to austrian vs. german telco (they do a few things differently here)
[14:26] <th> perhaps you were even on the wrong branch
[14:26] <th> i suggest getting chan_misdn from beronet and using the makefile to install kernel modules
[14:26] <th> and userspace
[14:28] <fake> i did
[14:28] <fake> even the mqueue branch
[14:28] <th> k
[14:28] <th> yes i'm on 0.3.0
[14:28] <th> with mq
[14:28] <th> i contributed for call deflection
[14:28] <th> and this works only mqueue
[14:28] <fake> i c
[14:28] <fake> weird.
[14:29] <fake> i guess the avmfritz driver is not as maintained as the hfcpci one
[14:30] <fake> and guess that's why i got kernel oopses yesterday, without even an L1 connection active ;)
[14:32] <th> i'm on btw
[14:33] <fake> oh
[14:35] <th> that's because of udev mainly
[14:35] Action: fake going to a technical museum now
[14:35] <fake> cya later ;)
[14:35] <th> cya
[15:00] <stf^rocklinux> [14:17:12] <th> can qemu use a block device as disk?
[15:01] <daja77> iirc yes, but it is damn slow
[15:01] <stf^rocklinux> th: according to the user manual, it should work to e.g. use /dev/hda as disk...
[15:02] <stf^rocklinux> or even /dev/hda1 ^^
[15:02] <daja77> i used the cdrom drive e.g.
[15:41] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) left irc: Remote closed the connection
[15:41] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) joined #rocklinux.
[15:41] <th> hmm ok
[15:42] <th> i just want to test 1st stage. so perhaps i just dont need any real block device
[15:43] <stf^rocklinux> th: the you can use the direct linux boot feature ;)
[15:43] <th> sounds neat
[15:43] <stf^rocklinux> th: use 'qemu -kernel <kernel-img> -initrd <initrd-img> -append <boot-opts>'
[15:43] <th> rocks
[15:44] <stf^rocklinux> th: you need to supply a (virtual) hard drive for some reason, though (with -hda)
[15:45] <stf^rocklinux> th: yeah, that's a nice little feature of qemu :)
[15:47] <th> kernel oopses
[15:47] <th> hmm
[15:51] <th> oh
[15:51] <th> QEMU PC emulator version 0.6.1
[16:00] <th> no oops with 0.7.2 :)
[16:06] <daja77> right old qemu oopsed on linux 2.6
[16:07] <th> now i get unable to mount root fs (my initrd propably)
[16:07] <th> strange
[16:07] <th> it's a cramfs
[16:08] <th> and the kernel is standard linux26  hmm
[16:08] <th> -append root=/dev/ram0  should be ok for initrd, no?
[16:09] <stf^rocklinux> th: yes, maybe try ext2fs as initrdfs?
[16:09] <th> i took everything from r7333 iso
[16:09] <th> kernel has cramfs support there
[16:09] <th> that's strange
[16:11] <th> unknown-block(0,0)
[16:11] <th> hmm
[16:11] <th> ah
[16:11] <stf^rocklinux> th: /dev/ram0 should be block(1,0)...
[16:11] <th> might be that qemu has a problem with our 16k initrd?
[16:12] <th> stf^rocklinux: yes - i was just trying /dev/ram/0 i had /dev/ram0 before all the time. and it's unknown-block(1,0) now.
[16:12] <stf^rocklinux> th: don't think so. I've got this error now and then testing stage1 with qemu...
[16:12] <th> "Couldn't find valid RAM disk iamge starting at 0"
[16:12] <stf^rocklinux> th: maybe qemu doesn't deal properly with compressed ram disks ...
[16:12] <th> i want to see the scollback before the panic
[16:12] <th> damn
[16:13] <stf^rocklinux> th: use -nographic (and sigkill to end qemu)
[16:13] <th> it's the kernel mounting it, no?
[16:13] <stf^rocklinux> th: yes
[16:13] <th> how does -nographic and sigkill help?
[16:13] <stf^rocklinux> th: if the ramdisk contains a proper fs, then it is at least mounted by the kernel
[16:14] <th> yes.
[16:14] <th> so you think it could help to gunzip it?
[16:14] <stf^rocklinux> th: -nographic uses text only output to console, so you can scroll back
[16:14] <th> with -nographic i get no output at all
[16:14] <th> no text
[16:15] <stf^rocklinux> th: not sure about that, I haven't pinned that bug down exactly
[16:15] [raphael] (n=raphael@raphael.netpark.at) left irc: Read error: 104 (Connection reset by peer)
[16:15] <stf^rocklinux> th: started from konsole?
[16:15] <th> xterm
[16:16] <th> -nographic      disable graphical output and redirect serial I/Os to console
[16:16] <th> hmm
[16:17] <stf^rocklinux> th: I think you need to -append console=ttyS0 with -nographic as well
[16:17] <stf^rocklinux> th: that worked here before
[16:18] <th> ok let's try
[16:18] <th> ttyS0?
[16:18] <th> that's serial
[16:18] <th> then i could use -serial to my tty
[16:18] <stf^rocklinux>  -nographic      disable graphical output and redirect serial I/Os to console
[16:18] <th> ah ok
[16:19] <th> stf^rocklinux: what's your qemu version?
[16:19] <th> 0.7.2?
[16:19] <stf^rocklinux> th: a limited 0.8.0, but currently I have none installed
[16:19] <stf^rocklinux> I'm trying to fix the 0.8.0 build atm.
[16:20] <th> cool
[16:20] <th> there was 0.9 already, no?
[16:20] <stf^rocklinux> th: not that I know of
[16:20] <th> oh no
[16:20] <th> confused
[16:23] [raphael] (n=raphael@raphael.netpark.at) joined #rocklinux.
[16:27] <th> this sucks
[16:30] <stf^rocklinux> th: -nographic requires a kernel with console on serial port enabled, I think that's why it doesn't work with crystal kernels
[16:30] <stf^rocklinux> th: did you try the uncompressed initrd? And did you check the initrd itself?
[16:32] <stf^rocklinux> th: btw. I assume you are trying to boot an x86 kernel and initrd, I've had strange errors with other architectures...
[16:42] <th> 16:32:33 < stf^rocklinux> th: -nographic requires a kernel with console on serial port 
[16:42] <th> ah yea
[16:42] <th> that should be the problem
[16:43] <th> yea i tried to uncomress..
[16:43] <th> stf^rocklinux: yea x86
[16:44] daja77_ (n=daja77@dslb-088-072-042-014.pools.arcor-ip.net) joined #rocklinux.
[16:44] <th> stf^rocklinux: we should enable serial console in crystal kernels
[16:45] <stf^rocklinux> th: and netlink as builtin module (for udev)
[16:45] <th> yea we have that
[16:45] <th> dont we?
[16:46] <stf^rocklinux> th: seems so
[16:54] daja77 (n=daja77@dslb-088-072-042-014.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out)
[17:00] <netrunner> th: fake is extremely against enabling serial console :)
[17:01] <netrunner> I also had this because it is useful for qemu remote testing
[17:06] ija_ (n=ija@ joined #rocklinux.
[17:09] ija (n=ija@ left irc: Nick collision from services.
[17:09] Nick change: ija_ -> ija
[17:58] <th> fake is against enabling serial console?
[17:59] <mnemoc> why?
[18:04] <netrunner> iirc because he has some hardware that has no serial console.
[18:05] <netrunner> I once sent a patch to do so but it was rejected due to this neg vote
[18:05] <mnemoc> doh
[18:09] <th> a kernel with serial support has problems on hardware without serial console?
[18:10] <th> i dont think so
[18:11] <th> checking if image is initramfs...it isn't (bad gzip magic numbers); looks like an initrd
[18:11] <th> Bad page state at free_hot_cold_page (in process 'swapper', page c100a300)
[18:11] <th> Backtrace:
[18:11] <th>  [<c0146880>] bad_page+0x80/0xc0
[18:11] <th>  [<c01470e3>] free_hot_cold_page+0x53/0x160
[18:11] <th>  [<c0113734>] free_initrd_mem+0x64/0x80
[18:11] <th>  [<c01002b0>] init+0x0/0x180
[18:11] <th> ...
[18:15] <stf^rocklinux> th: crystal rev. 7333 kernel and initrd?
[18:16] <th> yea
[18:16] <th> qemu
[18:16] <th> i bet that's the issue
[18:16] <th> cause i can boot them fine
[18:16] <th> (on non-qemu)
[18:17] <th> RAMDISK driver initialized: 16 RAM disks of 16384K size 1024 blocksize
[18:17] <th> perhaps this is a bit too much for 128MB qemu ram
[18:18] <stf^rocklinux> th: yeah, maybe qemu 0.8.0 works better
[18:18] <th> hmm same with 512m
[18:18] <netrunner> th: 16MB < 128MB imho
[18:18] <th> RAMDISK driver initialized: 16 RAM disks of 16384K size 1024 blocksize
[18:18] <th> 16 disks of 16384K size
[18:19] <th> 16 * 16384K
[18:19] <netrunner> oh.
[18:19] <th> that wont be allocated at once of course.
[18:19] <th> but might have mapping problems
[18:19] <th> i dont know
[18:19] <th> stf^rocklinux: got 0.8.0 working?
[18:20] <stf^rocklinux> th: not quite yet. A manual build finished successfully...
[18:21] <th> stf^rocklinux: that would be enough for me
[18:22] <stf^rocklinux> th: then you could use the --target configure option and leave out the targets that don't build
[18:23] <stf^rocklinux> (i386-user and mipsel-user, iirc)
[18:23] <th> the bad page is gone if i build the kernel with pentium2
[18:24] <th> still Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(1,0)
[18:24] <th> yea 
[18:24] <th> RAMDISK: Couldn't find valid RAM disk image starting at 0.
[18:25] <th> stf^rocklinux: if i leave out i386-user - isn't that exactly what i need?
[18:26] <stf^rocklinux> th: no, 'qemu' is still built. *-user are emulations that don't need root privileges
[18:27] <th> i need root privileges?
[18:27] <th> ./i386-user/qemu-i386
[18:27] <th> built fine
[18:29] <th> i'm so much confused today
[18:29] <stf^rocklinux> th: at least  for 'real' networking support you need root access
[18:29] <stf^rocklinux> th: not sure about details, though
[18:38] <th> QEMU PC emulator version 0.8.0, Copyright (c) 2003-2005 Fabrice Bellard
[18:38] <th> built fine
[18:39] <th> problem still not fixed. must be some other issue
[18:43] <th> stf^rocklinux: our 7333 mkinitrd is not able to create initramfs, or is it?
[18:43] <stf^rocklinux> th: not really. It seems cpio archives cannot store special files???
[18:43] <th> the code is in there
[18:44] <stf^rocklinux> th: I haven't looked at how the linux kernel does it yet ...
[18:44] <th> they propably dont have special files
[18:46] feistel (n=feistel@ joined #rocklinux.
[18:52] <th> there is something wrong at my site i guess.
[18:52] <th> i cant even get a ext2 initrd working
[18:54] <th> perhaps the 16k is just too big for qemu
[18:59] <mnemoc> .oO( my initrd is 2.3M, and works without trobules at qemu )o
[19:00] <th> so it's too small
[19:02] <th> -rw-r--r--  1 root root 1012M Mar 25 18:03 initrd-
[19:03] <th> i guess that's my /proc/core again
[19:03] <mnemoc> -rw-r--r--   1 root root 2,3M 2006-03-25 15:03 initrd-2.6.16-dist.img
[19:03] <mnemoc> ^---- cpio.gz
[19:04] <mnemoc> .oO( 1012M?? )o
[19:05] <stf^rocklinux> th: latest mkinitrd should be able to deal with special files ...
[19:05] <stf^rocklinux> th: oh, you used initramfs...
[19:18] <th> yea cpio. ack on:
[19:18] <th> 18:46:08 < stf^rocklinux> th: not really. It seems cpio archives cannot store special files???
[19:18] <th> bbl
[19:31] owl (n=owl@ left irc: "pissed."
[19:45] nookie_ (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) joined #rocklinux.
[19:48] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) left irc: Read error: 104 (Connection reset by peer)
[20:18] nookie_ (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) left irc: Read error: 110 (Connection timed out)
[20:19] <stf^rocklinux> th: qemu is fixed in https://www.rocklinux.net/submaster/smadm.cgi?i=2006032519463428193
[20:39] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) joined #rocklinux.
[20:45] feistel (n=feistel@ left irc: Remote closed the connection
[21:04] <th> stf^rocklinux: well - it built fine for me with ./configure && make && make install
[21:04] <stf^rocklinux> hmm...
[21:04] <th> stf^rocklinux: (2.0.3 stable host with gcc-3.2.3)
[21:04] <th> but cool anyways
[21:05] <stf^rocklinux> th: I'm using Crystal rev.7330
[21:05] <th> i need this as soon as my initrd thing works
[21:05] <th> and i'd like to work on this with qemu
[21:05] <th> i bet my best choice is to go for initramfs
[21:05] <stf^rocklinux> th: I didn't even get the same error messages with Emerge-Pkg and manual builds :S
[21:06] <th> but cpio obviously can't do special files
[21:06] <th> devices that is
[21:06] <th> so we need not to put them in but have a script create them
[21:06] <th> or something
[21:06] <stf^rocklinux> th: yes
[21:15] <th>   - The content of /lib/udev/devices directory which contains the nodes,
[21:15] <th>     symlinks and directories, which are always expected to be in /dev, should
[21:15] <th>     be copied over to the tmpfs mounted /dev, to provide the required nodes
[21:15] <th>     to initialize udev and continue booting.
[21:15] <th> hmmm
[21:15] <th> the devices shall be in there
[21:16] <th> it would be weird to have a script create them there so that we can copy them later
[21:17] <th> but hey we could have a tar in the cpio
[21:27] <stf^rocklinux> th: or an extra scipt that creates device nodes
[21:29] <th> is there a distribution already doing initramfs?
[21:30] <icelbox> geexbox?
[21:30] <icelbox> but I am unsure about that
[21:30] <th> a voice from the no-where
[21:30] <th> hi icelbox 
[21:31] <icelbox> hi ^^
[21:45] <th> http://lwn.net/Articles/157676/
[21:46] <th> see section "populating initramfs"
[21:46] <th> it tells about devices
[21:46] <th> in form of a text file
[21:46] <th> it's all done in the kernel source i think ;-/
[21:48] <th> gen_init_cpio (from kernel src) seems to be able to create a cpio containing devices
[21:58] owl (n=owl@ joined #rocklinux.
[21:59] <stf^rocklinux> th: we could reuse that program in rockinitrd...
[22:00] <th> yes
[22:00] <th> it's really simple
[22:00] <th> and it did not change since 2.6.12
[22:00] <th> still the question remains how to add such a cpio
[22:00] <th> giving it as cpio.gz as initrd= does not seem to be enough
[22:01] <stf^rocklinux> th: The kernel messages suggest otherwise 
[22:01] <stf^rocklinux> th: there's one like "checking if initrd is initramfs..."
[22:02] <owl> re
[22:02] <stf^rocklinux> wb owl
[22:03] <owl> thx
[22:03] <th> checking if image is initramfs...it isn't (bad gzip magic numbers); looks like an initrd
[22:03] <th> stf^rocklinux: guess what i passed
[22:04] <th> i used gen_init_cpio with a small list
[22:04] <th> then i manually gzipped it.
[22:04] <th> and gave it as initrd
[22:04] <th> perhaps one should not use gzip directly
[22:05] <th> The initramfs buffer format is based around the "newc" or "crc" CPIO
[22:05] <th> formats, and can be created with the cpio(1) utility.  The cpio
[22:05] <th> archive can be compressed using gzip(1).  One valid version of an
[22:05] <th> initramfs buffer is thus a single .cpio.gz file.
[22:05] <th> hmm
[22:05] <th> .   initramfs  := ("\0" | cpio_archive | cpio_gzip_archive)*
[22:05] <th> so even compressing is not necessart
[22:06] <th> perhaps i should stop using qemu
[22:06] <th> seems to f*ck up my passed initrd
[22:07] <stf^rocklinux> th: I haven't got initramfs to work so far ( neither with qemu nor otherwise)
[22:07] <th> have you tried?
[22:07] <stf^rocklinux> th: with qemu, yes
[22:07] <th> i bet it's a qemu problem
[22:07] <th> checking if image is initramfs...it isn't (bad gzip magic numbers); looks like an initrd
[22:07] <stf^rocklinux> th: but not with gen_init_cpio
[22:08] <th> this happens with the easiest hello world gen_init_cpio list
[22:21] <stf^rocklinux> th: isolinux/initrd.gz from the Crystal rev. 7333 ISO can be booted with qemu, you just have to use another kernel...
[22:22] <th> another kernel?
[22:22] <stf^rocklinux> th: I used my own kernel config
[22:22] <th> same here
[22:23] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) left irc: Read error: 104 (Connection reset by peer)
[22:23] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) joined #rocklinux.
[22:27] netrunner (n=andreas@anvame.net) left irc: Remote closed the connection
[22:31] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) left irc: Remote closed the connection
[22:32] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) joined #rocklinux.
[22:33] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) left irc: Remote closed the connection
[22:34] netrunner (n=andreas@anvame.net) joined #rocklinux.
[22:35] <th> stf^rocklinux: are you sure that passing irfs as initrd= through bootloader is really intended to work?
[22:35] <stf^rocklinux> th: no
[22:39] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) joined #rocklinux.
[22:40] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) left irc: Read error: 104 (Connection reset by peer)
[22:42] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) joined #rocklinux.
[22:52] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) left irc: Remote closed the connection
[22:52] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) joined #rocklinux.
[23:00] <th> i think it all comes down to: initramfs is not really supposed to be rebuilded as often as you change your socks
[23:00] <th> but anyways i'd appreciate some script to extract and implant irfs archive into kernel image
[23:00] <th> that's propably some ld hack
[23:01] nookie (n=nookie@85-124-48-125.dynamic.xdsl-line.inode.at) left irc: Read error: 104 (Connection reset by peer)
[23:06] <stf^rocklinux> th: yeah, such a script would be nice
[00:00] --- Sun Mar 26 2006