On Sat, Dec 29, 2012 at 6:39 PM, Milan Broz <gmazyland@xxxxxxxxx> wrote:
On 12/30/2012 12:15 AM, .. ink .. wrote:interesting, cryptsetup close (action_close) uses this and it doesn't fail here...
>
> i meant "tcrypt" for "truecrypt".
>
> Way to test.
> 1. create a truecrypt mapper.
> 2. call "crypt_init_by_name" to create context through the opened mapper and the function call will fail.
I think i have found the problem,here is a test to show it:
[root@mtz build]# cryptsetup open --type tcrypt /home/ink/src/truecrypt tcrypt
Enter passphrase:
[root@mtz build]# cryptsetup status /dev/mapper/tcrypt
/dev/mapper/tcrypt is active.
Device /dev/mapper/tcrypt_1 not found
[root@mtz build]# cryptsetup status tcrypt
/dev/mapper/tcrypt is active.
type: TCRYPT
cipher: aes-xts-plain64
keysize: 512 bits
device: /dev/loop0
loop: /home/ink/src/truecrypt
offset: 256 sectors
size: 40448 sectors
skipped: 256 sectors
mode: read/write
Enter passphrase:
[root@mtz build]# cryptsetup status /dev/mapper/tcrypt
/dev/mapper/tcrypt is active.
Device /dev/mapper/tcrypt_1 not found
[root@mtz build]# cryptsetup status tcrypt
/dev/mapper/tcrypt is active.
type: TCRYPT
cipher: aes-xts-plain64
keysize: 512 bits
device: /dev/loop0
loop: /home/ink/src/truecrypt
offset: 256 sectors
size: 40448 sectors
skipped: 256 sectors
mode: read/write
As the above two commands show with "status" argument,the command fail when the given path is full path.
"crypt_init_by_name" shows the same behavior,it fail when the path given is full path and pass if it is only the name of the mapper.
I have always used full path on the second argument with "crypt_init_by_name".
_______________________________________________ dm-crypt mailing list dm-crypt@xxxxxxxx http://www.saout.de/mailman/listinfo/dm-crypt
