I found it - it is a problem with ramdisk even when it is loaded not from
a floppy but in normal life. Behold:
# dd if=rootfs of=/dev/ram0 bs=1k count=1440
1440+0 records in
1440+0 records out
# mount -t minix /dev/ram0 /mnt
# df
Filesystem Type 1k-blocks Used Available Use% Mounted on
/dev/hda4 ext2 7542572 1680908 5478516 23% /
none proc 0 0 0 - /proc
/dev/hda1 ext2 23302 12734 9365 58% /boot
none devpts 0 0 0 - /dev/pts
/dev/ram0 minix 1421 1225 196 86% /mnt
# ll /mnt
total 15
drwxr-xr-x 11 root root 352 Nov 14 08:42 ./
drwxr-xr-x 23 root root 4096 Nov 13 20:13 ../
drwxr-xr-x 2 root root 96 Nov 15 16:14 bin/
drwxr-xr-x 2 root root 1696 Nov 15 20:33 dev/
drwxr-xr-x 2 root root 384 Nov 15 21:10 etc/
drwxr-xr-x 2 root root 192 Nov 15 16:15 lib/
drwxr-xr-x 2 root root 64 Nov 14 08:42 mnt/
drwxr-xr-x 2 root root 64 Nov 14 08:42 proc/
drwxr-xr-x 2 root root 128 Nov 15 16:14 sbin/
drwxr-xr-x 2 root root 64 Nov 14 08:42 tmp/
drwxr-xr-x 3 root root 96 Nov 14 08:55 usr/
# od -tx4 -Ax /mnt/sbin/init
000000 00000000 00000000 00000000 00000000
*
006420 00000000
006424
No wonder ELF loader refused to load and now it is clear where were we
getting all those zeroes in binfmt-0000 request.
So, it correctly gets
directory structure but not the file data blocks.
Investigation continues - at least I can debug it from the comfort of the
running kernel instead of the annoying reboots from the floppy etc.
Regards,
Tigran.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/