i always get a kernel panic:
Unable to handle kernel NULL pointer dereference at virtual
address 0000001c
current->tss.cr3 = 02cda000, %cr3 = 02cda000
*pde = 00000000
Oops: 0000
CPU:    0
EIP:    0010:[<c0124b7f>]
EFLAGS: 00010293
eax: 00000000   ebx: ffffffff   ecx: c37f6330   edx:
00001180
esi: c2cfdf80   edi: ffffff86   ebp: c3943d10   esp:
c2cfdf78
ds: 0018   es: 0018   ss: 0018
Process restore (pid: 60, process nr: 14,
stackpage=c2cfd000)
Stack: c3943d10 00000000 00000046 00001180 ffff0000
c012c6b3 c0e58000 c3943d10
       00000001 c2cfc000 c0124c49 c3943d10 00000000
00000000 c2cfc000 00000000
       0805cfb2 bfffed34 c01097f8 0805cfb2 00000000
00000000 00000000 0805cfb2
Call Trace: [<c012c6b3>] [<c0124c49>] [<c01097f8>]
Code: 8b 40 1c 85 c0 74 6e bb 00 e0 ff ff 21 e3 80 b9 88 00
00 00

when i want to restore my root-fs via:

(cd /mnt; zcat /root.dump.gz | restore -rvf-)
or
(cd /mnt; cat root.dump | restore -rvf-)

but not on:
(cd /mnt; restore -rvf root.dump)
or
(cd /mnt/home; zcat /home.dump.gz | restore -rvf-)


i use kernel 2.2.5-15, compiled with egcs-2.91.66, running
on an Celeron 466, 64MB ram

how can i restore without wasting space for the
uncompressed dump file???