Re: encryption

From: Stefan Kaczmarek <stefan_at_thezonie.org>
Date: Thu, 14 Jun 2007 15:53:46 -0700

Well, that's definitely a weird case.

On Jun 14, 2007, at 3:52 PM, Dylan Douglas wrote:

> Ivan says that he is running clients on the server to help with
> publishing. So, somehow it's getting confused and talking to the
> broker rather than the clients. I have some investigation to do.
>
> From: Stefan Kaczmarek [mailto:stefan_at_thezonie.org]
> Sent: Thursday, June 14, 2007 3:50 PM
> To: Dylan Douglas
> Cc: Ivan Kwok; Ben Ebert; Colin Keller; Gerald Rode; Jay Mairs;
> Nainesh Solanki; Sergio Alvarez; Ty Heath
> Subject: Re: encryption
>
> No, it sends it from whatever port it picks from the ether ... (I
> bind to 0)
>
> 8222 is the 0x201e port ... That's just the server's udp port.
>
> You could just ignore those messages ... I don't know why your ip
> and port would be in there, though.
>
> On Jun 14, 2007, at 2:50 PM, Dylan Douglas wrote:
>
>> It shouldn't be getting them on port 8222 though, should it? That
>> would be something that the would be on some other port since it
>> is p2p, not p2s.
>>
>> Z, does the p2p client always send udp traffic from the 8222 port?
>>
>> From: Ivan Kwok
>> Sent: Thursday, June 14, 2007 2:45 PM
>> To: Dylan Douglas
>> Cc: Stefan Kaczmarek; Ben Ebert; Colin Keller; Gerald Rode; Jay
>> Mairs; Nainesh Solanki; Sergio Alvarez; Ty Heath
>> Subject: Re: encryption
>>
>> You're probably seeing the app's or console's UDP messages. Don't
>> forget the console app and the client app are also running on the
>> miivi server for publishing stuff.
>>
>> -Ivan
>>
>>
>> On Jun 14, 2007, at 2:41 PM, Dylan Douglas wrote:
>>
>>> Okay now that I know I'm not on an acid trip, I'm seeing op code
>>> 0x10 and 0x20s and things like that. They are the right length
>>> for the type of packet they are supposed to be, but am I supposed
>>> to be handling these messages some how? Like passing them on or
>>> something? Cause the protocol spreadsheet says they are just p2p
>>> communications.
>>>
>>> From: Stefan Kaczmarek [mailto:stefan_at_thezonie.org]
>>> Sent: Thursday, June 14, 2007 12:45 PM
>>> To: Dylan Douglas
>>> Cc: Ben Ebert; Colin Keller; Gerald Rode; Ivan Kwok; Jay Mairs;
>>> Nainesh Solanki; Sergio Alvarez; Ty Heath
>>> Subject: Re: encryption
>>>
>>> I decrypted both of those packets fine. The second one was op
>>> code 2, length 0x1a, a sha and ip 65.120.42.225.
>>>
>>> I use the exact same code to encrypt and decrypt ... What do you
>>> need from me to figure this out?
>>>
>>> private static void DoCrypt(int key,byte[] data,int offset)
>>> {
>>> byte bite=0;
>>> for(int i=offset;i<data.length;i++)
>>> {
>>> // key = (key * 1103515245 + 12345) & 0x7fffffff;
>>>
>>> // Bork the key
>>> int a=key>>>16;
>>> int b=key&0xffff;
>>> a=(a * (1103515245 >>> 16))+12345;
>>> b=(b * (1103515245 & 0xffff))+12345;
>>> key=(key ^ a) ^ b;
>>>
>>> bite=(byte)((key/65536) % 256);
>>> data[i] ^= bite;
>>> }
>>> }
>>>
>>>
>>> - Z
>>>
>>> On Jun 14, 2007, at 12:32 PM, Dylan Douglas wrote:
>>>
>>>> Okay, finally got Wireshark to work on the MiiVi server. Here's
>>>> the
>>>> deal:
>>>>
>>>> I get the encrypted ping:
>>>> { 0x61, 0xef, 0xb9, 0x16, 0x75, 0x28, 0xb3 }
>>>> Decrypt it and get:
>>>> { 0x00, 0x00, 0x00 }
>>>> Yay teh shit works, right?
>>>>
>>>> I get the pierce, I'll assume by the length:
>>>> { 0xb8, 0xde, 0xa1, 0x23, 0xff, 0x0f, 0xc4, 0xd9, 0x51, 0x93, 0x8d,
>>>> 0x85, 0x3d, 0xe0, 0xa8, 0x87, 0x5a, 0xf6, 0x93, 0x69, 0xdc,
>>>> 0xdb, 0x8c,
>>>> 0x5d, 0x30, 0x2e, 0x1c, 0xfd, 0x8b, 0xcf, 0x8f, 0x2a, 0xf1 }
>>>> Decrypt it and get:
>>>> { 0x03, 0x01, 0x19, 0x0c ..... }
>>>> Doh! WTF is a 0x03 op code? Also the length is screwed, since
>>>> it should
>>>> come out to be 1A.
>>>>
>>>> Both of these are from my office machine, so why does one work
>>>> and the
>>>> other is borkered? I checked several ping packets and they all
>>>> work.
>>>> All the other packets are fucked. Can you check your encryption is
>>>> being happy for the piercing and not just the pinging? The
>>>> weird thing
>>>> is the packet aren't unencrypted either.
>>>>
>>>>
>>>>
>>>> -----
>>>> Dylan Douglas
>>>> MediaDefender
>>>>
>>>>
>>>>
>>>> <winmail.dat>
>>>
>>
>
Received on Fri Sep 14 2007 - 10:55:55 BST

This archive was generated by hypermail 2.2.0 : Sun Sep 16 2007 - 22:19:46 BST