[ntp:security] NOEPEER patch

Harlan Stenn stenn at nwtime.org
Fri Aug 3 10:48:38 UTC 2018


On 8/3/2018 3:46 AM, Martin Burnicki wrote:
> Harlan,
> Harlan Stenn wrote:
>> On 8/3/2018 2:28 AM, Martin Burnicki wrote:
>>> Harlan Stenn wrote:
>>>> and now it is:
>>>>   if (!peer || !peer->keyid || !(peer->flags & FLAG_SKEY)
>>>> 	return INVALIDNAK;
>>>> and I think we want:
>>>>   if (!peer || (!peer->keyid && !(peer->flags & FLAG_SKEY))
>>>> 	return INVALIDNAK;
>>> The fact that the FLAG_SKEY test has been in the code before Pearly's
>>> change doesn't necessarily mean that the test is correct, and required.
>>> Similar to the MODE_ACTIVE reply in this case instead of a MODE_PASSIVE
>>> reply that would be expected, even according to Dave.
>>> So *why* is FLAG_SKEY tested here? If a packet with an invalid/unknown
>>> key was received then it should make no difference if the *key* was a
>>> symmetric one, or an autokey one.
>> All I know right now (at 0237) is that this test dates back to August of
>> 2001, and possibly before that.
>> It *might* have something to do with the recollection I have that during
>> the first few packets of the autokey dance, there will not (yet) be a keyid.
> OK, can you eventually discuss this with DLM?


> I'm going to try if he change you proposed above will also result in the
> behavior expected in our case.

Much appreciated :)

> Martin

Harlan Stenn <stenn at nwtime.org>
http://networktimefoundation.org - be a member!

More information about the security mailing list