Hello there, As you can see on the reference to the forum, I've asked some help for this functionality but got no answer. So, i'm asking you if you'll find this helpful and would like to help me in making it working. The script is inside the post. Thanks in advance and best regards, Michael
What's the aim of this? If you want to inject extra files into the initrd or real_root just use a cdupdate.sh script that calls tar to decompress a tarball.
Well, i use gentoo for my livecd (www.pentoo.ch) and I thought that this could be useful for people who want to customize it, as well as people that want to customize the gentoo livecd. Decompressing tarballs use ram, and if somebody want to add many modules (like an entire e17 desktop module), using squashfs would be more apropriate, at least for people who doesn't have a lot of ram. This can also be helpful for minimal fix, as it would be possible to overlap a package present with a new version without having to rebuild the entire root squashfs image. Also the people who don't have a lot of bandwith for downloading a full iso can download just the module and modify themselves the iso to add it. I would understand that it may not be needed for gentoo livecd but if you have any advice for me to help solve my problem, so i would then try to contribute by providing patches for the genkernel init script and it may then interest you more, as you probably have other things to do atm.
And your problem goes away if you don't run the unionfs code? Or does it still remain?
Well, if i launch the livecd and it contains no modules, it doesn't crash. It's only when i load some .mo modules. I try this with the latest genkernel and unionfs 1.1.1 and kernel 2.6.14-r4. If you want, i can build a livecd to test this or i can chase you on irc so we can and i would try to explain it better.
Ok, i could make this work with unionfs 1.0.14 and some gen_compile.sh patching. For now, i successfully loaded one module and i could boot inside the livecd and launch the app. I'll test with several module loading and post my result. Would you like me to contribute a patch for genkernel to support this?
Tim? Any comments?
Hi, I've made 3 patches that i'll attach right after this. I did not tested it with genkernel-4 as i have not much time atm, but it should work. The code is not that nice, feel free to enhance it. Cheers
Created attachment 88986 [details, diff] busybox-1.00-rt-mdstart.plasmaroo-max_loop.patch I made a patch against this version but it seems fixed in 1.1 now according to : http://www.busybox.net/cgi-bin/viewcvs.cgi/trunk/busybox/libbb/loop.c?rev=11099&r1=10933&r2=11099
Created attachment 88987 [details, diff] linux-2.6.16-gentoo-r9-max_loop.patch This one allow more than 8 loop devices
Created attachment 88988 [details, diff] genkernel-4.patch
None of these even applies to current SVN, and it doesn't seem like anybody cares anymore since there's not been any activity on this bug for over a year. Please reopen if you come up with something that works with current SVN and doesn't involve patching the kernel and busybox.
