You are not logged in.

#26 2005-12-29 23:13:37

phrakture
Arch Overlord
From: behind you
Registered: 2003-10-29
Posts: 7,879
Website

Re: ArchCK - Current status and initrd/initramfs progress.

dibblethewrecker wrote:

You can call the initrd flyingbananakungfumonkey.rat

Have you been looking at my menu.lst?

Offline

#27 2005-12-30 00:17:29

lilsirecho
Veteran
Registered: 2003-10-24
Posts: 5,000

Re: ArchCK - Current status and initrd/initramfs progress.

Recent posts indicate obfuscation rather than definitive coding to establish the correct LILO entry for the initramfs system. This assumes that an entry for that option is actually required for an initramfs boot which seems a hard item to get definitive information for on this forum.

You can call the initrd flyingbananakungfumonkey.rat and put it anywhere you like (* as long as it is available at boot) as long as it it correctly ref'ed in the bootloader (Quote DTW)

It does indeed have to be in the bootloader and in proper coded format in order that the system boots.

In fact, this user suspects the original initrd code will boot the system, and thusly, it boots into initrd and not initramfs unless ....some argument is made during boot to change to initramfs.  Is such the case?

How does the user determine the system is using initramfs as opposed to initrd after the system is booted?

=============================================================
A question no one has answered...When is a zero always equal to a one?


Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit!     X-ray confirms Iam spineless!

Offline

#28 2005-12-30 05:17:54

iphitus
Forum Fellow
From: Melbourne, Australia
Registered: 2004-10-09
Posts: 4,927

Re: ArchCK - Current status and initrd/initramfs progress.

The user doesnt need to know.

Soon all the user will need to know that if they are running ArchCK, they use mkinitramfs, and edit /etc/mkinitramfs.conf. Standard Arch Kernel is the same.

The kernel needs to know, and it can identify between the two formats, a set size ext2 filesystem initramfs, or a gzipped cpio image initramfs.

As we've defined, the lilo/grub entry can be anything, so long as it is accessible by the bootloader, with lilo, thats when you run 'lilo'. and with grub, thats at boot time, so best to be safe and keep it in /boot despite the fact it can go anywhere and be named practically anything. Including flyingbananakungfumonkey.rat

Which i think is an awesome name.

iphitus

Offline

#29 2005-12-30 16:29:29

dtw
Forum Fellow
From: UK
Registered: 2004-08-03
Posts: 4,439
Website

Re: ArchCK - Current status and initrd/initramfs progress.

lilsirecho wrote:

Recent posts indicate obfuscation rather than definitive coding...

"obfuscate" - is that like saying things in a manner more complicated than they needs to be?

Offline

#30 2005-12-30 16:50:59

phrakture
Arch Overlord
From: behind you
Registered: 2003-10-29
Posts: 7,879
Website

Re: ArchCK - Current status and initrd/initramfs progress.

lilsirecho wrote:

Recent posts indicate obfuscation rather than definitive coding to establish the correct LILO entry for the initramfs system. This assumes that an entry for that option is actually required for an initramfs boot which seems a hard item to get definitive information for on this forum.

Well, there's a huge difference between telling a user exactly what to do, and how to do it for themselves.  The point is that it doesn't matter what you name the initramfs, as long as the bootloader line points to the file.  It's exactly like any other file.
I think what everyone is trying to do here is not to obfuscate things, but to try to make the user understand that it's just a file and can be na,ed whatever you want it to.  If you don't know the name, do an 'ls'... it's that simple.

lilsirecho wrote:

In fact, this user suspects the original initrd code will boot the system, and thusly, it boots into initrd and not initramfs unless ....some argument is made during boot to change to initramfs.  Is such the case?

Nope, the kernel does the detection - in fact, it loads the initrd image, determines if it is a ext2 filesystem (initrd) or a cpio image (initramfs) and performs what it needs to accordingly.

lilsirecho wrote:

How does the user determine the system is using initramfs as opposed to initrd after the system is booted?

You don't and it doesn't matter... the only way to determine it would be to check the type of the image that you used in the initrd line.  If it's a ext2 filesystem, it's an initrd, if it's a cpio image, it's an initramfs.

On a side note:
iphitus, using pivot_root in an initramfs is not recommended and can cause issues.  If you are using this, I'd suggest getting/building klibc and using the static "run-init" app from there in the initramfs.

Offline

#31 2005-12-30 23:43:29

iphitus
Forum Fellow
From: Melbourne, Australia
Registered: 2004-10-09
Posts: 4,927

Re: ArchCK - Current status and initrd/initramfs progress.

On a side note:
iphitus, using pivot_root in an initramfs is not recommended and can cause issues.  If you are using this, I'd suggest getting/building klibc and using the static "run-init" app from there in the initramfs.[/quote]

pivot root is whats used by the Arch initrd setup. Is it that its safe to use in initrd? or we're both 'wrong'?

Can you give me a link to where you found this info?

Offline

#32 2005-12-30 23:51:15

phrakture
Arch Overlord
From: behind you
Registered: 2003-10-29
Posts: 7,879
Website

Re: ArchCK - Current status and initrd/initramfs progress.

iphitus wrote:

pivot root is whats used by the Arch initrd setup. Is it that its safe to use in initrd? or we're both 'wrong'?

Can you give me a link to where you found this info?

Right, that's why I brought it up, it's fine in initrd, but because initramfs is not a real mounted filesystem, some differences show up.

IIRC I read that in the header comment for run-init in the klibc source tarball.

Later when I get a chance, I'll add klibc to [community].

Offline

#33 2005-12-31 00:25:07

dtw
Forum Fellow
From: UK
Registered: 2004-08-03
Posts: 4,439
Website

Re: ArchCK - Current status and initrd/initramfs progress.

klibc is in my repo, phrak, please take it from there if you don't have your own version:

http://dtw.jiwe.org/pkgs/dibble/PKGBUIL … c/PKGBUILD

obviously you can remove the depends line and use `uname -r`...

maybe I should have done that a while ago?  Well, no, cos my pkg DOES depend on kernel26arch smile

What will yours depend on?

Offline

#34 2005-12-31 04:18:29

iphitus
Forum Fellow
From: Melbourne, Australia
Registered: 2004-10-09
Posts: 4,927

Re: ArchCK - Current status and initrd/initramfs progress.

So whats the difference? how do i implement run-init..

Looks like im going to have to install klibc inside the initramfs which will soo not be fun

Offline

#35 2005-12-31 06:54:29

phrakture
Arch Overlord
From: behind you
Registered: 2003-10-29
Posts: 7,879
Website

Re: ArchCK - Current status and initrd/initramfs progress.

iphitus wrote:

Looks like im going to have to install klibc inside the initramfs which will soo not be fun

Why not? You just replace the 'cp /blah/busybox' with 'cp /usr/lib/klibc/bin/*' (path might be wrong) and 'cp /lib/klibc-*.so'

klibc supplies most of the tools you'll need during boot, and you save about 400K (uncompressed).

dibblethewrecker wrote:

klibc is in my repo, phrak, please take it from there if you don't have your own version

Heh, didn't know that - I made my own - no big deal, but mine does not copy the klibc Kbuild files to the kernel - that's not needed yet.

Offline

#36 2005-12-31 06:55:40

phrakture
Arch Overlord
From: behind you
Registered: 2003-10-29
Posts: 7,879
Website

Re: ArchCK - Current status and initrd/initramfs progress.

iphitus wrote:

So whats the difference? how do i implement run-init..

The klibc stuff don't supply --help stuff (it'd be dumb if they did), so you have to view the comments in the source files.  It's farly self explanatory.

Offline

#37 2005-12-31 16:00:08

dtw
Forum Fellow
From: UK
Registered: 2004-08-03
Posts: 4,439
Website

Re: ArchCK - Current status and initrd/initramfs progress.

phrakture wrote:

Heh, didn't know that - I made my own - no big deal, but mine does not copy the klibc Kbuild files to the kernel - that's not needed yet.

Au contraire, I may be wrong (I often am) but I believe you need those bits to build splashutils, for gensplash, properly - so a pkg in [community] of klibc would do well to include them smile

Offline

#38 2006-01-02 21:55:10

lilsirecho
Veteran
Registered: 2003-10-24
Posts: 5,000

Re: ArchCK - Current status and initrd/initramfs progress.

iphitus;

Appreciate your patient reply to my post and am happy tp report that I have a third kernel installed and running with archck7.

Thanks for a zippy OS!


Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit!     X-ray confirms Iam spineless!

Offline

#39 2006-01-04 00:26:11

Stalwart
Member
From: Latvia, Riga
Registered: 2005-10-18
Posts: 445
Website

Re: ArchCK - Current status and initrd/initramfs progress.

When will 2.6.15-archck be relesed? big_smile


IRC: Stalwart @ FreeNode
Skype ID: thestalwart
WeeChat-devel nightly packages for i686

Offline

#40 2006-01-08 18:19:07

ozar
Member
From: USA
Registered: 2005-02-18
Posts: 1,686

Re: ArchCK - Current status and initrd/initramfs progress.

Stalwart wrote:

When will 2.6.15-archck be relesed? big_smile

Soon, I hope because I'd really like to try it, too.  The new stock 2.6.15-2 kernel breaks ALSA on my box (and I've not found a fix), but maybe it'll still work with this kernel in place.


oz

Offline

#41 2006-01-11 19:54:12

nuopus
Member
From: Mesa, AZ
Registered: 2005-03-09
Posts: 60

Re: ArchCK - Current status and initrd/initramfs progress.

ozar wrote:
Stalwart wrote:

When will 2.6.15-archck be relesed? big_smile

Soon, I hope because I'd really like to try it, too.  The new stock 2.6.15-2 kernel breaks ALSA on my box (and I've not found a fix), but maybe it'll still work with this kernel in place.

And hopefully it doesnt give me the same problems as the stock 2.6.15 did with the new ati drivers. X hard locks when it exits using the 8.16.20 fglrx drivers from ATI using the stock 2.6.15. I read that some other users were having the same problems as well. Using the archck 2.6.14 kernel works NICELY with those drivers.

Someone suggested that there may be an incompatibility with the 2.6.15 kernel? If so i'm hoping that ati fixes the issue before archck updates their kernel. Any insight on this?

Offline

#42 2006-01-24 23:32:21

Bengi
Member
Registered: 2005-10-25
Posts: 21

Re: ArchCK - Current status and initrd/initramfs progress.

Hi,

I was reading through the post and I noticed that phrakture suggested that using pivot_root in initramfs can cause issues. I am posting because I have just done a pacman -Syu  and upgraded to the latest archck kernel. I have previously been using a custom archck kernel, which does not use initramfs or intrd. When I changed over to initramfs and the latest kernel my md device nodes were not creaed in /dev. I backtracked through the init scripts and the problem seems to be that when pivot_root executes from the init script in the initramfs image it comes up with Invalid argument. This means that the device nodes from the initramfs image don't get copied to /dev and I can't start my raid arrays. I have gone back to using a custom kernel, however I would quite like to use the initramfs method and I suspect others may be having the same issue.

I take it that the reason for this (as pointed out by phrakture), is that unlike initrd, initramfs is not a real mounted file system and pivot_root has problems with this.

If this is the case then will the kernel26archck package in community be updated to use klibc and run-init instead of busybox and pivot_root?

Thanks in advance

Offline

#43 2006-01-25 14:41:45

iphitus
Forum Fellow
From: Melbourne, Australia
Registered: 2004-10-09
Posts: 4,927

Re: ArchCK - Current status and initrd/initramfs progress.

ignore the pivot root line, I need to remove it. It has nothing to do with your raid array.

No, your raid is not working because you havnt configured it in /etc/initramfs.conf, and then generated a new initramfs.

Please also ensure that as well as configuring your initramfs properly, you define root as /dev/md0 not /dev/md/0.

iphitus

Offline

#44 2006-01-25 22:54:17

Bengi
Member
Registered: 2005-10-25
Posts: 21

Re: ArchCK - Current status and initrd/initramfs progress.

I have edited the initramfs.conf, enabled raid, configured the RAID_ROOT_ARRAY and RAID_ROOT_DEVICES variables correctly and generated a new initramfs. I have also made sure the syntax is correct. The problem with the raid is that only the md device node that is configured in initramfs.conf by RAID_ROOT_ARRAY is created. This means that mdadm is unable to start the other 4 arrays. In rc.sysinit on line 45 rc.sysinit checks to see if /initrd/dev exists and if it does then tries to copy /initrd/dev/md? to /dev. On my system I think that the reason that the device nodes are not created is that /initrd/dev doesn't exist when rc.sysinit executes and rc.sysinit is therefore unable to copy the device nodes. I may be wrong, but if I uncompress the initramfs image and look at the init script, I can't see any other point other than pivot_root where it remounts / as new_root/initrd. If this is the case then once chrooted from the initramfs image to my root filesystem /initrd/dev doesn't exist and rc.sysinit can't copy of the device nodes.

Offline

#45 2006-03-07 10:45:50

Bengi
Member
Registered: 2005-10-25
Posts: 21

Re: ArchCK - Current status and initrd/initramfs progress.

Hello,

I haven't heard anything regarding this in a while. I was just wondering if this has been fixed yet or if anybody knows when it might be fixed.

Thanks in advance.

Offline

#46 2006-03-07 11:20:10

iphitus
Forum Fellow
From: Melbourne, Australia
Registered: 2004-10-09
Posts: 4,927

Re: ArchCK - Current status and initrd/initramfs progress.

try the latest release,

if that doesnt work, try mkinitrd and vanilla. if that doesnt work, file a bug.

iphitus

Offline

#47 2006-03-07 11:40:50

Bengi
Member
Registered: 2005-10-25
Posts: 21

Re: ArchCK - Current status and initrd/initramfs progress.

Thanks iphitus, I am not in front of that box right now (it is my home pc), but on the archlinux box I am using (my work pc), I have just installed the latest mkinitramfs to see if there was any reference to run-init in /sbin/mkinitramfs but there is not. I am also pretty sure that the normal archlinux kernel will work because pivot_root remounts root as new_root and rc.sysinit can copy my device nodes and therefore start my arrays. I will however test this when I get a chance and if the normal archlinux kernel works and archck fails file a bug report.

Offline

#48 2006-03-15 19:58:26

napoleon
Member
Registered: 2005-04-03
Posts: 176

Re: ArchCK - Current status and initrd/initramfs progress.

I just upgraded to the latest kernel26archck 2.6.15.archck7-2.1 and rebooted. But my nice frambuffer/splash is gone. My fonts are all big and ugly.

I didn't change anything kust rebooted  tongue


A good friend will come and bail you out of jail...BUT a true friend will be sitting next to you saying, "Damn...that was fun!"

Offline

#49 2006-03-18 01:33:07

Bengi
Member
Registered: 2005-10-25
Posts: 21

Re: ArchCK - Current status and initrd/initramfs progress.

I have just tried the latest release and it did not work. It has the same problem as before, but in mkinitramfs the line containing pivot_root has been commented out so the error message about pivot_root is gone. I also tried the latest vanilla and it worked.

I have filed a bug report here:-

http://bugs.archlinux.org/task/4188

Offline

#50 2006-03-18 13:14:52

gradgrind
Member
From: Germany
Registered: 2005-10-06
Posts: 921

Re: ArchCK - Current status and initrd/initramfs progress.

Bengi:
You seem to be having a sad time with your RAID array.
I have a question: If you are really set on using archck, why don't you just do a little edit on rc.sysinit and add 'mknod' commands for your missing devices (I assume you know, or can find out, what they are)? At least until a more permanent fix can be thought up.

It seems to be quite possible to use run-init to clean up the initramfs (my larch live CD uses it - but without the rest of klibc, I compiled run-init.c separately, linking with uclibc), but I imagine it also cleans up the device nodes in the initramfs, so that after it exits, they are not there any more to copy out.

It would be possible to change the initramfs code to copy the device nodes out, but I suspect that wouldn't be a good idea (consider people who use a custom kernel without initramfs). In other words I suspect the real fix lies in rc.sysinit and not in initramfs or initrd.

Offline

Board footer

Powered by FluxBB