Hi Javier,

Thanks for the information.
I will start digging in to issue with wireshark and get back to the group.
 Quick question on mesh secured vs mesh non-secured.

If i created a non-secured mesh first then decided to secure the mesh
afterwords without
Taking down the non-secured mesh (adding it while active in non_secure
mode.) . Will that effect the authsae output and or behavior.  If authsae
creats the link for the mesh with channel and id then wouldnt it cancel out
the non_secure mesh anyways.

Thanks

Fred

On Wednesday, October 17, 2012, Javier Cardona <[email protected]> wrote:
> Hi Fred,
>
> Thanks for providing the logs.  Some observations:
>
> 1. The "broken" frame reported in dmesg is from 00:15:6d:94:6e:6b to
> 00:15:6d:94:59:d8 and is not found in the authsae logs.  There are no
> frames to that destination in the authsae logs.
>
> 2. The "broken" frame is an unsecured peer link open frame (category
> 13 = 0xd = IEEE80211_CATEGORY_MESH_ACTION).  See the >><< below
>
> [68981.912493] XXX: frame: 00000000: d0 00 00 00 00 15 6d 94 59 d8 00
> 15 6d 94 6e 6b  ......m.Y...m.nk
> [68981.912499] XXX: frame: 00000010: 00 15 6d 94 6e 6b 00 00>>0d<<01
> 84 0f 1f 01        ..m.nk........
>
> 3. authsae only sends *secured* peer link management frames (category
> 15 = 0xf = IEEE80211_CATEGORY_SELF_PROTECTED).  For instance:
>
> tx frame hexdump
> d0 00 00 00 00 15 6d 94 4f 76 00 15 6d 94 6e 6b 00 15 6d 94
> 6e 6b 00 00>>0f<<01 10 00 01 08 82 84 8b 96 8c 12 98 24 32 04
>
> 4. If that peer link open frame did not originate in authsae, it must
> have originated in the kernel.  In fact, it is the kernel that
> establishes peer links in unsecured (open) mesh networks.  But that
> protocol is disabled once a secure mesh is started.
>
> 5. Or maybe it was not started?
>
>> ----------
>> meshd: Starting mesh with mesh id = my-mesh900
>> ----------
>> basic rates: hexdump
>> 82 84 8b 96 8c 98 b0
>> ----------
>>
>> Unexpected error -16 (expected -107)
>
> Error -16 is NLE_SEQ_MISMATCH, so it appears as if the command to
> start the secured mesh has failed.  Apparently authsae ignores that
> error code and proceeds thinking a secure mesh has successfully
> started.  But the kernel seems to be running its own unsecured mesh
> under the hood.
>
> So far this is only a hypothesis.  To be sure we would have to see the
> beacons from the node that raises all the errors:  do those beacons
> have a Robust Secure Network Information element in them?  (The RSN IE
> looks like:
> const char rsn_ie[0x16] = {0x30, /* RSN element ID */
>                        0x14, /* length */
>                        ...
> and is correctly parsed by wireshark).
>
> If they don't, we would have confirmed the hypothesis.
>
> Thanks for your help on this!
>
> Javier
>
>
>
>
> On Tue, Oct 16, 2012 at 10:19 PM, fred veldini <[email protected]>
wrote:
>> Hello here is a dump of the authsae mesh errors with a dmesg dump.
>> Hope this helps
>>
>> Fred
>>
>> Things that standout are the following
>>
>> debug in authsae:
>> NL80211_CMD_TX_STATUS (1136259849.264671)
>> tx frame failed!tx_frame(0x80612a0, 0xbfdf06ec, 128)
>> tx frame (seq num=0)
>>
>>
>> dmesg output:
>>  ath: phy0: Failed to stop TX DMA, queues=0x005!
>>  ath: phy0: Timeout while waiting for nf to load:
AR_PHY_AGC_CONTROL=0x40d22
>>
>>
>> Raw output of authsae and dmesg below
>>
>>
>> ----------------------------------
>> authsae start cut
>> ----------------------------------
>> # setting freq 2432, mode 0 (seq num=1136259852)
>>
>> ----------
>> nlcfg rates hexdump
>> 0a 00 14 00 37 00 6e 00 3c 00 5a 00 78 00 b4 00 f0 00 68 01
>> e0 01 1c 02
>> ----------
>>
>> ----------
>> mgtk:  hexdump
>> 19 a1 95 d6 07 0a 84 37 09 d6 54 c2 ed 76 ba ee
>> ----------
>>
>> ----------
>> Fixed Information Elements in this STA hexdump
>> 01 08 82 84 8b 96 8c 12 98 24 32 04 b0 48 60 6c
>> ----------
>>
>> meshd: Starting mesh with mesh id = my-mesh900
>> ----------
>> basic rates: hexdump
>> 82 84 8b 96 8c 98 b0
>> ----------
>>
>> Unexpected error -16 (expected -107)
>> NL80211_CMD_NEW_PEER_CANDIDATE(1136259846.237476)
>> tx_frame(0x80612a0, 0xbfdf001c, 128)
>> tx frame (seq num=1136259863)
>> ----------
>> tx frame hexdump
>> b0 00 00 00 00 15 6d 94 5a 39 00 15 6d 94 6e 6b 00 15 6d 94
>> 5a 39 00 00 03 00 01 00 00 00 13 00 49 f2 86 1a 2f b3 31 47
>> 92 57 d5 2f 35 fb 00 d9 e1 15 de 14 46 05 18 fd ac 3f fc a9
>> 10 7f a8 dd b8 0a 15 a1 74 be 0b ad 0f 8b 07 18 7c 12 64 95
>> 5d 41 c8 cf 1b 91 c2 6b 06 4a 48 90 1c a2 b3 db f0 32 eb 7f
>> aa f1 2e eb c5 6b cc 47 43 63 76 e7 8f 46 84 db 5d 76 1b 89
>> 4e f3 e4 3e 38 2e db 95
>> ----------
>>
>> new unauthed sta (seq num=1136259864)
>> NL80211_CMD_NEW_STATION (1136259846.252940)
>> NL80211_CMD_TX_STATUS (1136259846.256306)
>> tx frame failed!NL80211_CMD_NEW_PEER_CANDIDATE(1136259846.442116)
>> tx_frame(0x80612a0, 0xbfdf001c, 128)
>> tx frame (seq num=1136259865)
>> ----------
>> tx frame hexdump
>> b0 00 00 00 00 15 6d 94 4f 76 00 15 6d 94 6e 6b 00 15 6d 94
>> 4f 76 00 00 03 00 01 00 00 00 13 00 7b d3 ca a4 93 55 f4 f1
>> ce 86 61 39 15 90 a4 ef e5 cd cf ea 6c a1 c3 6e de 73 d9 0f
>> bc cd 62 a3 7d b8 b0 60 8a de 5c be 07 18 8a 9b de ad 57 b7
>> 5e d2 87 01 33 51 b3 09 38 52 07 4b 22 b6 30 1d 7f 84 60 28
>> b1 76 96 70 a9 0f 52 6b fc e4 a0 67 ff ba 3d 74 4e 38 ed 8d
>> 1d c7 07 43 a3 68 5a 37
>> ----------
>>
>> new unauthed sta (seq num=1136259866)
>> NL80211_CMD_TX_STATUS (1136259846.458128)
>> NL80211_CMD_NEW_STATION (1136259846.458361)
>> NL80211_CMD_FRAME (1136259846.461391)
>> ----------
>> rx frame hexdump
>> b0 00 3a 01 00 15 6d 94 6e 6b 00 15 6d 94 4f 76 00 15 6d 94
>> 6e 6b 80 c5 03 00 01 00 00 00 13 00 e8 21 ee 50 26 bb 83 81
>> d1 9c 37 3c e6 1a 48 ba f6 07 97 9c 09 28 5f 5e 87 b7 9c 59
>> 70 0b c9 f3 ca 15 a6 14 e5 b3 45 b8 6d c3 08 69 11 ad b4 8a
>> 3f df df 71 a2 5d 0e 46 09 f0 76 90 bc 21 a6 4d dd 6d 2e a3
>> 3e 91 89 ac 8e c1 ce 17 bb f3 4c 98 eb ac 60 61 ca 5f 53 83
>> 70 37 1a 2a f2 ed b2 b5
>> ----------
>>
>> tx_frame(0x80612a0, 0xbfdf02ac, 64)
>> tx frame (seq num=1136259867)
>> ----------
>> tx frame hexdump
>> b0 00 00 00 00 15 6d 94 4f 76 00 15 6d 94 6e 6b 00 15 6d 94
>> 4f 76 00 00 03 00 02 00 00 00 01 00 80 86 92 5a c4 a3 25 4a
>> 82 cd 44 77 c4 84 00 d6 8c 4d 16 58 dc 83 6e 4c bb 39 98 e1
>> 15 ce f3 41
>> ----------
>>
>> NL80211_CMD_FRAME (1136259846.483637)
>> ----------
>> rx frame hexdump
>> b0 00 3a 01 00 15 6d 94 6e 6b 00 15 6d 94 4f 76 00 15 6d 94
>> 6e 6b 90 c5 03 00 02 00 00 00 01 00 e4 28 2b de 83 a9 6e 1f
>> c0 bc ed 15 bb fb 68 50 4a 5c 9d 25 19 fd 51 3e f6 50 bf c5
>> 8d de 6d e3
>> ----------
>>
>> fin: 0, key len:32 peer:00:15:6d:94:4f:76 me:00:15:6d:94:6e:6b
>> ----------
>> pmk hexdump
>> 32 11 49 4d 2e 55 85 59 1e d5 99 4d de 4d 90 54 34 0a 0f f2
>> 77 41 65 b7 3b 6b--
> Javier Cardona
> cozybit Inc.
> http://www.cozybit.com
> _______________________________________________
> Devel mailing list
> [email protected]
> http://lists.open80211s.org/cgi-bin/mailman/listinfo/devel
>
_______________________________________________
Devel mailing list
[email protected]
http://lists.open80211s.org/cgi-bin/mailman/listinfo/devel

Reply via email to