On 7/20/21 10:28 PM, Vladislav Bolkhovitin wrote:
>
> On 7/18/21 3:21 PM, Hannes Reinecke wrote:
>> On 7/17/21 9:22 AM, Sagi Grimberg wrote:
>>>> Implement NVMe-oF In-Band authentication. This patch adds two new
>>>> fabric options 'dhchap_key' to specify the PSK
>>>
>>> pre-shared-key.
>>>
>>> Also, we need a sysfs knob to rotate the key that will trigger
>>> re-authentication or even a simple controller(s-plural) reset, so this
>>> should go beyond just the connection string.
>>>
>>
>> Yeah, re-authentication currently is not implemented. I first wanted to
>> get this patchset out such that we can settle on the userspace interface
>> (both from host and target).
>> I'll have to think on how we should handle authentication; one of the
>> really interesting cases would be when one malicious admin will _just_
>> send a 'negotiate' command to the controller. As per spec the controller
>> will be waiting for an 'authentication receive' command to send a
>> 'challenge' payload back to the host. But that will never come, so as it
>> stands currently the controller is required to abort the connection.
>> Not very nice.
>
> Yes, in this case after some reasonable timeout (I would suggest 10-15
> seconds) the controller expected to abort connection and clean up all
> allocated resources.
>
> To handle DoS possibility to make too many such "orphan" negotiations,
> hence consume all controller memory, some additional handling is needed.
> For simplicity as a first step I would suggest to have a global limit on
> number of currently being authenticated connections.
>
> [...]
>
>>>> + chap->key = nvme_auth_extract_secret(ctrl->opts->dhchap_secret,
>>>> + &chap->key_len);
>>>> + if (IS_ERR(chap->key)) {
>>>> + ret = PTR_ERR(chap->key);
>>>> + chap->key = NULL;
>>>> + return ret;
>>>> + }
>>>> +
>>>> + if (key_hash == 0)
>>>> + return 0;
>>>> +
>>>> + hmac_name = nvme_auth_hmac_name(key_hash);
>>>> + if (!hmac_name) {
>>>> + pr_debug("Invalid key hash id %d\n", key_hash);
>>>> + return -EKEYREJECTED;
>>>> + }
>>>
>>> Why does the user influence the hmac used? isn't that is driven
>>> by the susbsystem?
>>>
>>> I don't think that the user should choose in this level.
>>>
>>
>> That is another weirdness of the spec.
>> The _secret_ will be hashed with a specific function, and that function
>> is stated in the transport representation.
>> (Cf section "DH-HMAC-CHAP Security Requirements").
>> This is _not_ the hash function used by the authentication itself, which
>> will be selected by the protocol.
>> So it's not the user here, but rather the transport specification of the
>> key which selects the hash algorithm.
>
> Yes, good catch. It looks as a minor errata material to specify that
> hash function here is implementation specific.
>
> I would suggest to just hardcode SHA512 here. Users don't have to be
> confused by this.
>
Sure, can do. My reasoning was that the target absolutely has to support
the hash functions specified in the PSK, so that will be a safe bet to
choose for the hash function in the protocol itself.
(Any other hash function _might_ not be preset on the target.)
But if the PSK does not specify a hash the target need to pick one; and
for that of course we can use SHA512.
Cheers,
Hannes
--
Dr. Hannes Reinecke Kernel Storage Architect
hare@xxxxxxx +49 911 74053 688
SUSE Software Solutions Germany GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), GF: Felix Imendörffer
