markus reichelt | 24 Sep 2008 19:16
Picon

Re: loop-aes: problem after repartitioning disk

* Juergen Lemke <lemke_juergen <at> yahoo.de> wrote:

> Until yesterday I had a hard drive with three partitions, one used
> as a root partition encrypted with loop-aes, another unencrypted
> partition and swap space. I had to delete the unencrypted partition
> and divide it into two smaller, separate partitions, which I did
> using cfdisk. As a result, cfdisk shows that the partition
> encrypted with loop-aes changed names from /dev/hda3 to /dev/hda4.

Shouldn't be a problem. I successfully migrated encrypted systems to
different partitions.

> After this change I could not boot my encrypted system from the
> usual usb stick. I booted using a knoppix cdrom, mounted my usb key
> under /boot and edited initrd.conf on the usb key so that it read
> 
> CRYPTROOT=/dev/hda4
> 
> (rather than /dev/hda3) and then ran
> 
> $ sudo ./build-initrd.sh /boot/initrd.conf [this copied a bunch of
> files and only generated warnings about not being able to preserve
> permissions on two files]

which ones?

> $ sudo umount /boot
> $ sudo syslinux /dev/sdc
> 
> But this did not work. When I now try to boot from the usb stick, I
(Continue reading)

Juergen Lemke | 24 Sep 2008 21:17
Picon
Favicon

Re: loop-aes: problem after repartitioning disk

Hi Markus,

Thank you for your help.

> > $ sudo ./build-initrd.sh /boot/initrd.conf [this
> copied a bunch of
> > files and only generated warnings about not being able
> to preserve
> > permissions on two files]
> 
> which ones?

Here is what I get once I run build-initrd.sh (after having edited build-initrd.sh to point
DESTINATIONPREFIX to the mount point of my usb key rather than /boot):

 knoppix <at> Knoppix:/ramdisk/home/knoppix/loop-AES-v3.2c$ sudo ./build-initrd.sh ../ek/initrd.conf
Loading config from '../ek/initrd.conf'
16+0 records in
16+0 records out
16384 bytes (16 kB) copied, 6.8158e-05 seconds, 240 MB/s
32 inodes
16 blocks
Firstdatazone=5 (5)
Zonesize=1024
Maxsize=268966912

Filesystem           1K-blocks      Used Available Use% Mounted on
/ramdisk/home/knoppix/loop-AES-v3.2c/tmp-i-900
                            11        10         1  91% /ramdisk/home/knoppix/loop-AES-v3.2c/tmp-d-900
-rw------- 1 root root 2440 Sep 24 14:59 /ramdisk/home/knoppix/ek/initrd.gz
(Continue reading)

markus reichelt | 25 Sep 2008 21:42
Picon

Re: loop-aes on Intel Mac?

* Juergen Lemke <lemke_juergen <at> yahoo.de> wrote:

> Mini-question: should I expect trouble trying to run Linux from a
> loop-aes encrypted root partition on a recent Intel-based Mac?

If it runs Linux, you are fine. 

--

-- 
left blank, right bald
Matthias Kaehlcke | 29 Sep 2008 11:24

loop-AES: symbol loop_register_transfer not found

hi,

i encountered a problem with loop-AES v3.2c on a ARM device with a linux 
2.6.20 kernel. the module builds successfully, but when i try to load it 
the kernel complains about a NULL pointer:

Unable to handle kernel NULL pointer dereference at virtual address 00000004
pgd = c0650000
[00000004] *pgd=0558b031, *pte=00000000, *ppte=00000000
Internal error: Oops: 17 [#1]
Modules linked in: audioctl keypad lirc_ep93xx leds spicodec
CPU: 0
PC is at lookup_symbol+0x38/0x78
LR is at vprintk+0x344/0x400
pc : [<c0057764>]    lr : [<c003a770>]    Not tainted
sp : c555beac  ip : c555be24  fp : c555bec4
r10: 00000001  r9 : c555bf78  r8 : 00000000
r7 : c555bf74  r6 : f8138000  r5 : bf034758  r4 : 00000000
r3 : 00000000  r2 : c555a000  r1 : bf034758  r0 : 0000002c
Flags: nzcv  IRQs on  FIQs on  Mode SVC_32  Segment user
Control: C000717F
Table: 00650000  DAC: 00000015
Process exe (pid: 1186, stack limit = 0xc555a250)
Stack: (0xc555beac to 0xc555c000)
bea0:                            bf025380 bf034758 00000000 c555bef0 
c555bec8
bec0: c00579c4 c005773

i put traces in the lookup_symbol() function of the kernel, the 
offending symbol seems to be loop_register_transfer:
(Continue reading)

Matthias Kaehlcke | 30 Sep 2008 13:10

Re: loop-AES: symbol loop_register_transfer not found

Jari Ruusu wrote:
> Matthias Kaehlcke wrote:
>> i encountered a problem with loop-AES v3.2c on a ARM device with a linux
>> 2.6.20 kernel. the module builds successfully, but when i try to load it
>> the kernel complains about a NULL pointer:
> [snip]
>> any idea what could be the problem?
> 
> My guess would be: somehow miscompiled module.

it's strange, it works on a very similar hardware with a slightly 
different kernel (same version, but different config).

> Can you try applying the kernel patch version of loop-AES (kernel-2.6.*.diff
> file in loop-AES tarball) to your kernel sources, and compile loop-AES
> built-in (not a module). Does that work?

the device isn't in my hands anymore, so i haven't been able to try it :(

if i get the opportunity i'll give it a try and report the results

meanwhile i'll work with a configuration where loop-AES works perfectly

thanks for your time!

Attachment (mkaehlcke.vcf): text/x-vcard, 336 bytes

Gmane