Re: cryptsetup not working under 3.6 - regression from 3.4?

From: Milan Broz
Date: Sun Oct 21 2012 - 15:29:22 EST


On 10/21/2012 02:36 PM, Tvrtko Ursulin wrote:
> On 21/10/12 13:20, Zdenek Kaspar wrote:
>>>> I would say you are still missing some modules.
>>>>
>>>>> Kernel says this:
>>>>> device-mapper: table: 252:1: crypt: Error allocating crypto tfm
>>>>> device-mapper: ioctl: error adding target to table
>>>>
>>>> It complains about aes-cbc-essiv:sha256.
>>>>
>>>> It can be missing CBC od SHA256, but according the message I bet
>>>> you have no "cbc" block cipher mode module compiled.
>>>>
>>>> Can you grep your final .config for CONFIG_CRYPTO_CBC and
>>>> CONFIG_CRYPTO_SHA256 a paste it here?
>>>
>>> I both working 3.4 and non-working 3.6 situation is the same:
>>>
>>> CONFIG_CRYPTO_CBC=y
>>> CONFIG_CRYPTO_SHA256=m
>>
>> Compare please:
>> grep CONFIG_CRYPTO /boot/config-3.4
>> grep CONFIG_CRYPTO /boot/config-3.6
>>
>> One of the problem could be that your configuration misses something
>> like: CONFIG_CRYPTO_BLKCIPHER, CONFIG_CRYPTO_MANAGER, etc.. or some of
>> those could changed into modules and are not getting loaded..
>
> Here it is:

Hm, so it should work without problem. Can you paste full output of failing
cryptsetup command (with added --debug switch) and your full kernel .config?

Cryptsetup itself has regression tests which test almost all common
combinations of ciphers so the problem is almost surely in some kernel
part misconfiguration.
Any other related messages in syslog beside two lines you posted?

Thanks,
Milan
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/