svn commit: r225748 - user/adrian/if_ath_tx/sys/dev/ath

[Adrian Chadd]

Author: adrian
Date: Sun Sep 25 07:38:33 2011
New Revision: 225748
URL: http://svn.freebsd.org/changeset/base/225748

Log:
  Update things

Modified:
  user/adrian/if_ath_tx/sys/dev/ath/README-crypto

Modified: user/adrian/if_ath_tx/sys/dev/ath/README-crypto
==============================================================================
--- user/adrian/if_ath_tx/sys/dev/ath/README-crypto	Sun Sep 25 02:41:50 2011	(r225747)
+++ user/adrian/if_ath_tx/sys/dev/ath/README-crypto	Sun Sep 25 07:38:33 2011	(r225748)
@@ -6,15 +6,6 @@ iperf tcp - client on STA, server on box
 
 ccmp_decap: whoa there, talk about whacked.
 ccmp_decap: ic_flags=0x000c2408
-ccmp_decap: tid=16, keyix=5, wk_rxkeyix=5
-TODS 00:23:6c:bf:38:3e->00:1b:b1:58:f6:f0(00:1b:b1:58:f6:f0) data WEP [IV a6 3e 00 e4 17 00 KID 0]
- 0841 3c00 001b b158 f6f0 0023 6cbf 383e 001b b158 f6f0 2015 a63e 0020 e417 0000 aaaa 0300 0000 888e 0103 005f 0203 0200 0000 0000 0000 0000 cb00 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0
-000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0069 4b5f 5689 d8de 3fda d3e2 7be3 ceaa 8b00 0079 ced4 2296 60bc ac
-
-.. then later:
-
-ccmp_decap: whoa there, talk about whacked.
-ccmp_decap: ic_flags=0x000c2408
 ccmp_decap: tid=0, keyix=5, wk_rxkeyix=5
 TODS 00:23:6c:bf:38:3e->00:1c:c0:43:87:2b(00:1b:b1:58:f6:f0) data QoS [TID 0] WEP [IV 88 09 00 11 19 00 KID 0]
  8849 2c00 001b b158 f6f0 0023 6cbf 383e 001c c043 872b 308b 0000 3000 8809 0020 1119 0066 d9a3 41da 6979 a898 3a05 495f 3233 3435 3637 3839 3031 3233 3435 3637 3839 3031 3233 3435 3637 3839 3031 3233 3435 3
@@ -31,15 +22,6 @@ TODS 00:23:6c:bf:38:3e->00:1c:c0:43:87:2
 wlan0: [00:23:6c:bf:38:3e] discard MPDU frame, BA win <2228:2291> (4 frames) rxseq 2227 tid 0 (retransmit)
 wlan0: [00:23:6c:bf:38:3e] discard MPDU frame, BA win <2236:2299> (3 frames) rxseq 2235 tid 0 (retransmit)
 
-.. then later:
-
-ccmp_decap: whoa there, talk about whacked.
-ccmp_decap: ic_flags=0x000c2408
-ccmp_decap: tid=16, keyix=5, wk_rxkeyix=5
-TODS 00:23:6c:bf:38:3e->00:1b:b1:58:f6:f0(00:1b:b1:58:f6:f0) data WEP [IV 2a 0a 00 11 19 00 KID 0]
- 0841 3c00 001b b158 f6f0 0023 6cbf 383e 001b b158 f6f0 601b 2a0a 0020 1119 0000 aaaa 0300 0000 888e 0103 005f 0203 0200 0000 0000 0000 0000 d700 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0
-000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 00af 015a 18af 9806 7015 56e0 a174 c423 d000 0098 73cc c4a6 13d9 99
-
 .. at which point the traffic had stopped hard; the RX side was discarding
 frames due to failing CCMP replay.
 
@@ -54,3 +36,23 @@ wlan0: [00:23:6c:bf:38:3e] AES-CCM repla
 
 So the question here - why is it that the crypto IV suddenly was invalid
 for one frame? Is it an RX or TX problem?
+
+Ie, the IV in the header:
+
+8809 0020 1119 0066
+
+.. and then what looks like some TCP payload from iperf ('0' -> '9' repeated?)
+with an 8 byte trailer.
+
+The IV sequence values:
+
+cynthia:crypto adrian$ ./de-iv.pl 112150606580104
+88 09 00 11 19 00 66
+cynthia:crypto adrian$ ./de-iv.pl 420547143
+47 0a 00 11 19 00 00
+
+There aren't enough bytes between the 802.11 header / IV and the beginning
+of the payload sequence for it to be an TCP header and if it were an
+IP frame, it'd have a valid looking IP frame version /header length.
+
+