Description of problem: The linux kernel provided with ipodlinux can't boot from an ext2 filesystem created with Fedora's mke2fs command. Version-Release number of selected component (if applicable): e2fsprogs-1.38-12, e2fsprogs-1.38-0.FC4.1 How reproducible: Always. Steps to Reproduce: 1. Plug in a 60Gig 5G (video) ipod 2. mke2fs /dev/sdb3 3. Actual results: It looks like it works, until you reboot, then you get these errors: VFS: Mounted root (ext2 filesystem). Mounted devfs on /dev Freeing init memory: 64K attempt to access beyond end of device 03:03: rw=0, want=2037609827, limit=128520 attempt to access beyond end of device 03:03: rw=0, want=2037609827, limit=128520 attempt to access beyond end of device 03:03: rw=0, want=1852400244, limit=128520 attempt to access beyond end of device 03:03: rw=0, want=1852400244, limit=128520 attempt to access beyond end of device 03:03: rw=0, want=1852400244, limit=128520 Kernel panic: No init found. Try passing init= option to kernel. This happens with 'mke2fs /dev/sdb3' and 'mke2fs -r 0 /dev/sdb3'. Using mke2fs from a live-cd distribution (Slax KillBill) (http://slax.linux-live.org/download.php) works. Expected results: No error, and a successful boot? Additional info: I've been frustrated for days by this bug. Other people with the same version of mke2fs and ext2fs library on Gentoo and Debian don't have this problem. See http://ipodlinux.org/forums/viewtopic.php?p=105478#105478
In short, has RedHat made incompatible changes to mke2fs?
We have made no incompatible changes that I am aware of. We _have_ switched mke2fs to enable online_resize by default, but that's a compatible change. That's not Red Hat-specific, either: the code is already in upstream e2fsprogs and has just been made the default for upstream mke2fs in e2fsprogs-1.39, too, so other distros will be picking up that default also. Does "mke2fs -O ^resize_inode" help?
Sorry, since getting it working, I haven't been motivated enough to reformat and start over.
Weird. The erroneous offsets, in hex, are ascii. attempt to access beyond end of device 03:03: rw=0, want=2037609827, limit=128520 obase=16 2037609827 79737563 Liberator:~ sandeen$ echo -n "ysuc" | hexdump -C 00000000 79 73 75 63 |ysuc| attempt to access beyond end of device 03:03: rw=0, want=1852400244, limit=128520 obase=16 1852400244 6E696274 Liberator:~ sandeen$ echo -n "nibt" | hexdump -C 00000000 6e 69 62 74 |nibt| it would have been nice if the 'attempt to access' message had given us a backtrace of how we got there.
I don't have an ipod I can test this with. Either someone can send me one ;-) or if anyone runs into this again, please let me know by re-opening this bug. (Or if the original reporter ever re-tests...) Until then, I don't think there's anything I can do with the (little) information I have here. -Eric