bcm43xx-d80211 broadcast reception with WPA
Subsystems (3)
Hi,
Long time lurker, first time poster. ^_^
I've been backporting the bcm43xx-d80211 driver to whatever the released 2.6 kernel was using the rt2x00 project's d80211 stack (equivalent to current wireless-dev but with a workaround for not having a ieee80211_dev pointer and still using the _tfm interface instead of the _cypher interface.)
As of last night's wireless-dev tree bcm43xx, everything seems to be operating fine except incoming broadcast traffic is coming in 14 bytes too long and scrambled. I presume this means it's not decrypting properly...
Anyway, I just thought I'd mention it. It might have gone unnoticed by the bcm43xx-d80211 developers, since it doesn't interfere with normal operation (A DHCP client's only broadcasts are outgoing) and only showed up for me because radvd's RAs were not arriving and my IPv6 address was not being set.
I couldn't find any mention of such a thing on the list, and I'm happy to provide whatever debugging output is useful, but the laptop with the device isn't with me at the moment.
Relevant facts:
Platform: Debian/unstable (PPC) w/linux-image-2.6.18-1-powerpc (2.6.18-3) Drivers: bcm43xx-d80211 from wireless-dev 774f233b7915a2c36480eb4d98e6f57938f04b7b Firmware: 4.80.46.0 (BE, from AppleAirPortBrcm4311) Stack: ieee80211 from http://rt2x00.serialmonkey.com/rt2x00-cvs-daily.tar.gz
2006110303 is the date on the output, I believe. Hasn't been updated since 20061028 Plus a backport of the following commits:
[PATCH] d80211: extend extra_hdr_room to be a bytecount 522e078b9f1f8309770dd161d90ddac1573a7877 [PATCH] d80211: remove unused variable in ieee80211_rx_irqsafe 10bfc9cdf9621385a3b69aa35f9fa86cc6a46bc6 [PATCH] d80211: Add wireless statistics 448bf25bc9e3d70a211fdf235426472089371c43
(as well as anything else that showed up in a diff of the d80211 dir against the rt2x00 iee80211 dir and wasn't a 2.6.19ism or wireless-devism)
I'm basically using the instructions I posted at [1] except also patching rt2x00's ieee80211 stack.
I acknowledge that any of the firmware version, the backporting, the forward porting or the current lunar cycle may be causing this problem. If no one pipes up with an insight, I'll try tonight with a v3 firmware, although the reason I moved to a v3 firmware was my previous build of bcm43xx-d80211 also wasn't getting an IPv6 address, although I don't believe the RAs were scrambled in that case.
Signature · 3 lines
Paul "TBBle" Hampson Opinions expressed here do not reflect the views of my employer Hell, we don't even agree on my pay cheque
Hi,
Quoted 4 lines
I've been backporting the bcm43xx-d80211 driver to whatever the released 2.6 kernel was using the rt2x00 project's d80211 stack (equivalent to current wireless-dev but with a workaround for not having a
ieee80211_devpointer and still using the_tfminterface instead of the_cypherinterface.)
Just a note about that stack, it should only be used on kernel 2.6.17 and higher.
Previous kernels had a bug that caused the stack to crash.
The _tfm cipher interface will soon change, since I haven't had time to update the stack to the latest version yet, I'll do that this weekend.
Ivo
Quoted 5 lines · Ivo Van Doorn <ivdoorn <at> gmail.com> writes:
Quoted 2 lines
I've been backporting the bcm43xx-d80211 driver to whatever the released using the rt2x00 project's d80211 stack
Just a note about that stack, it should only be used on kernel 2.6.17 and higher.
Thanks, I've noted that at the berlios wiki.
--
Paul "TBBle" Hampson
[email protected]
Quoted 4 lines · Paul Hampson
Hi,
Long time lurker, first time poster. ^_^
Quoted 8 lines
I've been backporting the bcm43xx-d80211 driver to whatever the released 2.6 kernel was using the rt2x00 project's d80211 stack (equivalent to current wireless-dev but with a workaround for not having a
ieee80211_devpointer and still using the_tfminterface instead of the_cypherinterface.)As of last night's wireless-dev tree bcm43xx, everything seems to be operating fine except incoming broadcast traffic is coming in 14 bytes too long and scrambled. I presume this means it's not decrypting properly...
It sounds like a bug in the hardware decryption setup. Are you using TKIP or not?
Incoming mcast frames are handled in a special way in hardware. The keyidx field of the packet is used to lookup the key, as far as I know. (Otherwise the MAC address is used).
Can I see a full dmesg log and a capture log on the broken machine, please?
Signature · 1 line
Greetings Michael.
Quoted 6 lines · Michael Buesch <mb <at> bu3sch.de> writes:
Quoted 4 lines · Paul Hampson
I've been backporting the bcm43xx-d80211 driver to whatever the released 2.6 kernel was using the rt2x00 project's d80211 stack (equivalent to current wireless-dev but with a workaround for not having a
ieee80211_devpointer and still using the_tfminterface instead of the_cypherinterface.)
Quoted 6 lines
Quoted 3 lines
As of last night's wireless-dev tree bcm43xx, everything seems to be operating fine except incoming broadcast traffic is coming in 14 bytes too long and scrambled. I presume this means it's not decrypting properly...
It sounds like a bug in the hardware decryption setup. Are you using TKIP or not?
Yes, it's using TKIP. The router docs and the loading of the tkip module when I use the softmac driver agree on this.
Quoted 4 lines
Incoming mcast frames are handled in a special way in hardware. The keyidx field of the packet is used to lookup the key, as far as I know. (Otherwise the MAC address is used).
Can I see a full dmesg log and a capture log on the broken machine, please?
Sending first some ipv6 broadcast pings and then some ipv4 broadcast pings:
(Commands "ping -b 192.168.192.255 -c 1" and "ping6 -I intel0 -c 1 ip6-allnodes")
17:04:08.794400 00:02:a5:40:8d:61 (oui Unknown) Unknown SSAP 0x26 > 33:33:00:00:00:01 (oui Unknown) Unknown DSAP 0x9c Unnumbered, eb, Flags [Poll], length 116
0x0000: 9d26 fbda 7284 bd60 6cdf 58c4 d064 71c6
0x0010: 2a09 adab 4a19 a691 5640 9216 eae8 df70
0x0020: b94e 9ee9 fd75 6c25 ab16 36fb cdac c231
0x0030: 0f17 f965 4d20 4a11 ab50 c77f 66ca 54e4
0x0040: e469 e458 5d6c c13d cc78 48fd da5c 7f71
0x0050: 2f06 0728 c8da 689b b790 ec22 4d62 5a92
0x0060: 221b b3cb 65e5 dc8a 8e57 3486 2a1e a2c2
0x0070: faf6 ae71
17:04:10.104111 00:02:a5:40:8d:61 (oui Unknown) Unknown SSAP 0xa8 >
33:33:00:00:00:01 (oui Unknown) Unknown DSAP 0xb8 Supervisory, Reject, rcv seq
78, Flags [Final], length 116
0x0000: b8a9 099d 0afb f9f3 8ef6 1c31 81c0 f1eb
0x0010: 3869 1952 9762 f4f0 c743 7613 fd9c 99cd
0x0020: 1644 a454 df14 5481 a35a 8c96 59b3 9391
0x0030: 8579 a165 3da2 58ad a6a8 d499 e40c 4c4c
0x0040: 3b33 a4ce 2b2e 439b 77f6 da5d 1d18 1685
0x0050: 1e64 39cb 3565 5596 30eb fa1c 1cbd cec8
0x0060: 395b a38c f7a4 a754 7c19 d694 a94b a4f9
0x0070: 5785 64aa
17:04:10.938418 00:02:a5:40:8d:61 (oui Unknown) Unknown SSAP 0x1c >
33:33:00:00:00:01 (oui Unknown) Unknown DSAP 0x64 Unnumbered, 23, Flags [Poll],
length 116
0x0000: 641c 338d 7f62 2bcd 4fff c7dd 4a6e efa1
0x0010: 07ed f39b 5b88 c68e 27dd f35b 70cf df3c
0x0020: 0cb8 f3ba 0b97 9069 43f9 e74f 1cb2 e4d7
0x0030: bf97 fbd8 94d8 efc5 284d 5393 604b 13ef
0x0040: 1cd7 46e1 7b35 008b 8247 89bb 0a6a 4dac
0x0050: 45e3 49af 853d 13fa e263 dea8 26ca 7076
0x0060: bec6 938f bd75 19bd a3ea 9f79 ea65 2c2d
0x0070: a45c b3d1
17:04:14.491735 00:02:a5:40:8d:61 (oui Unknown) Unknown SSAP 0x8a > Broadcast
Unknown DSAP 0xe4 Information, send seq 49, rcv seq 14, Flags [Final], length 96
0x0000: e58b 621d 383b 5114 c37d 54de da9e dd8b
0x0010: 7d28 87d2 7d53 cd57 f0b4 c079 54a5 0bbe
0x0020: 3eb2 f0b9 e1e6 e82b e52b ffaf f833 217c
0x0030: dbe7 c9f1 db0f 592e b432 5f7d 8041 f73f
0x0040: 7267 662b d64e 170d c619 a447 b2c0 bd17
0x0050: b97b b032 dd1b d8f5 c007 eae9 0aee ea9f
17:04:16.489911 00:02:a5:40:8d:61 (oui Unknown) Unknown SSAP 0xa4 > Broadcast
ProWay NM Information, send seq 122, rcv seq 28, Flags [Final], length 96
0x0000: 0fa5 f439 af38 57a6 564c 0c25 e2c0 7a09
0x0010: 61fb a1f1 adb0 3718 cb39 3a03 6ecf ad42
0x0020: 6e9c 75d7 cd06 0566 30c9 0238 4cf8 61a9
0x0030: 0928 9414 f48b 2a07 3eca 05de a8a9 9787
0x0040: 1ed5 2643 f82a b9a8 8e5a 5410 6858 b5c0
0x0050: ecb2 72a3 2dfb 66ac 0ce8 c4f8 ea87 ab14
17:04:18.449142 00:02:a5:40:8d:61 (oui Unknown) Unknown SSAP 0xb4 > Broadcast
Unknown DSAP 0x6a Supervisory, Receiver not Ready, rcv seq 23, Flags [Command],
length 96
0x0000: 6bb4 252e 2888 d3b1 68bc 6129 9087 4170
0x0010: f4e2 4a47 adc9 9bce 7bf2 51b3 ac20 bd10
0x0020: 7d67 3e00 6b6f 41ff 3e0c 2940 d31c 6143
0x0030: f7e4 caa6 879f 4663 e04b 0f6d 37eb 1db5
0x0040: fffd 0dfa 9b78 80e1 a30a 799e 9b1a 9d4a
0x0050: 61c8 f041 e564 d566 b697 aaf6 8336 a6f3And the dmesg output:
bcm43xx_d80211: no version for "ssb_init" found: kernel tainted.
PCI: Enabling device 0001:10:12.0 (0004 -> 0006)
ssb: Core 0 found: cc 0800, rev 04, vendor 4243
ssb: Core 1 found: cc 0812, rev 05, vendor 4243
ssb: Core 2 found: cc 080D, rev 02, vendor 4243
ssb: Core 3 found: cc 0807, rev 02, vendor 4243
ssb: Core 4 found: cc 0804, rev 09, vendor 4243bcm43xx_d80211: Broadcom 4306 WLAN found
ssb: Switching to core 4
ssb: Switching to core 1bcm43xx_d80211: PHY connectedbcm43xx_d80211: Detected PHY: Version: 2, Type 2, Revision 2bcm43xx_d80211: Detected Radio: ID: 2205017f (Manuf: 17f Ver: 2050 Rev: 2)bcm43xx_d80211: Radio turned offbcm43xx_d80211: Radio turned off
wmaster0: Selected rate control algorithm 'simple'bcm43xx_d80211: Virtual interface added (type: 0x00000002, ID: 6, MAC:
00:0d:93:ef:57:2d)
ssb: Switching to core 0
ssb: Switching to core 1bcm43xx_d80211: PHY connectedbcm43xx_d80211: firmware revision 15F, patchlevel 7E, date 2006-07-29 05:54:02
ssb: Switching to core 0
ssb: Switching to core 1bcm43xx_d80211: Radio turned on
ssb: Switching to core 0
ssb: Switching to core 1bcm43xx_d80211: Chip initializedbcm43xx_d80211: 30-bit DMA initializedbcm43xx_d80211: Keys clearedbcm43xx_d80211: Selected 802.11 core (phytype 2)
wlan0.11: Does not support passive scan, disabled
wlan0: starting scan
wlan0: scan completed
wlan0: Initial auth_alg=0
wlan0: authenticate with AP 00:14:6c:ed:c1:76
wlan0: RX authentication from 00:14:6c:ed:c1:76 (alg=0 transaction=2 status=0)
wlan0: authenticated
wlan0: associate with AP 00:14:6c:ed:c1:76
wlan0: RX AssocResp from 00:14:6c:ed:c1:76 (capab=0x471 status=0 aid=1)
wlan0: associatedbcm43xx_d80211: Using software based encryption for keyidx: 0, mac:
00:14:6c:ed:c1:76bcm43xx_d80211: Using software based encryption for keyidx: 1, mac:
ff:ff:ff:ff:ff:ff
agpgart: Putting AGP V2 device at 0000:00:0b.0 into 4x mode
agpgart: Putting AGP V2 device at 0000:00:10.0 into 4x mode
[drm] Loading R300 Microcode
wlan0: no IPv6 routers present
device wlan0 entered promiscuous mode
audit(1163225045.337:6): dev=wlan0 prom=256 old_prom=0 auid=4294967295
wlan0: TKIP decrypt failed for RX frame from 00:14:6c:ed:c1:76 (res=-3)
wlan0: TKIP decrypt failed for RX frame from 00:14:6c:ed:c1:76 (res=-3) device wlan0 left promiscuous mode
audit(1163225061.805:7): dev=wlan0 prom=0 old_prom=256 auid=4294967295
I can't reliably reproduce those "TKIP decrypt failed", I suspect that might be annoyingly co-incidental failures from SSH, NMB, DHCP or IPv6 RA traffic.
Unloading the driver dmesg:bcm43xx_d80211: Radio turned off
ssb: Switching to core 0bcm43xx_d80211: DMA-32 0x0200 (RX) max used slots: 2/64
ssb: Switching to core 1bcm43xx_d80211: DMA-32 0x02A0 (TX) max used slots: 0/128bcm43xx_d80211: DMA-32 0x0280 (TX) max used slots: 0/128bcm43xx_d80211: DMA-32 0x0260 (TX) max used slots: 0/128bcm43xx_d80211: DMA-32 0x0240 (TX) max used slots: 0/128bcm43xx_d80211: DMA-32 0x0220 (TX) max used slots: 20/128bcm43xx_d80211: DMA-32 0x0200 (TX) max used slots: 0/128bcm43xx_d80211: Virtual interface removed (type: 0x00000002, ID: 6, MAC:
00:0d:93:ef:57:2d)
wlan0: deauthenticate(reason=3)
wlan0: set_encrypt - low-level disable failed
wlan0: set_encrypt - low-level disable failed
wlan0: Initial auth_alg=0
wlan0: authenticate with AP 00:00:00:00:00:00
And loading up with v3 firmware:
ssb: Core 0 found: cc 0800, rev 04, vendor 4243
ssb: Core 1 found: cc 0812, rev 05, vendor 4243
ssb: Core 2 found: cc 080D, rev 02, vendor 4243
ssb: Core 3 found: cc 0807, rev 02, vendor 4243
ssb: Core 4 found: cc 0804, rev 09, vendor 4243bcm43xx_d80211: Broadcom 4306 WLAN found
ssb: Switching to core 4
ssb: Switching to core 1bcm43xx_d80211: PHY connectedbcm43xx_d80211: Detected PHY: Version: 2, Type 2, Revision 2bcm43xx_d80211: Detected Radio: ID: 2205017f (Manuf: 17f Ver: 2050 Rev: 2)bcm43xx_d80211: Radio turned offbcm43xx_d80211: Radio turned off
wmaster0: Selected rate control algorithm 'simple'bcm43xx_d80211: Virtual interface added (type: 0x00000002, ID: 8, MAC:
00:0d:93:ef:57:2d)
ssb: Switching to core 0
ssb: Switching to core 1bcm43xx_d80211: PHY connectedbcm43xx_d80211: firmware revision 122, patchlevel 9A, date 2005-08-15 18:44:03
ssb: Switching to core 0
ssb: Switching to core 1bcm43xx_d80211: Radio turned on
ssb: Switching to core 0
ssb: Switching to core 1bcm43xx_d80211: Chip initializedbcm43xx_d80211: 30-bit DMA initializedbcm43xx_d80211: Keys clearedbcm43xx_d80211: Selected 802.11 core (phytype 2)
wlan0.11: Does not support passive scan, disabled
wlan0: starting scan
wlan0: scan completed
wlan0: Initial auth_alg=0
wlan0: authenticate with AP 00:14:6c:ed:c1:76
wlan0: RX authentication from 00:14:6c:ed:c1:76 (alg=0 transaction=2 status=0)
wlan0: authenticated
wlan0: associate with AP 00:14:6c:ed:c1:76
wlan0: RX AssocResp from 00:14:6c:ed:c1:76 (capab=0x471 status=0 aid=1)
wlan0: associatedbcm43xx_d80211: Using software based encryption for keyidx: 0, mac:
00:14:6c:ed:c1:76bcm43xx_d80211: Using software based encryption for keyidx: 2, mac:
ff:ff:ff:ff:ff:ff
One each of the ip4 and ip6 pings under v3 firmware 17:18:29.454625 00:02:a5:40:8d:61 (oui Unknown) Unknown SSAP 0xdc > Broadcast Unknown DSAP 0xa6 Unnumbered, 07, Flags [Poll], length 96
0x0000: a6dc 1759 5095 ed4f 92f6 17dc d10c 9538
0x0010: e3e5 e302 e8f0 bfd8 0970 3dc7 315a c5eb
0x0020: ab6c 0c17 76a8 69ff c316 4955 1762 7ca0
0x0030: ba7f e65f e490 57e4 ad6c 53d5 fd4e d6de
0x0040: 41b2 5ab9 4749 52e4 1a9d bad9 e2a7 8544
0x0050: 91a6 eeef 5cc4 958c bc83 d7af 31f6 09a3
17:18:44.738765 00:02:a5:40:8d:61 (oui Unknown) Unknown SSAP 0xd2 >
33:33:00:00:00:01 (oui Unknown) Unknown DSAP 0x0a Information, send seq 103, rcv
seq 95, Flags [Command], length 116
0x0000: 0ad2 cebe 311a edd5 4e6c 3c2f 2e53 5120
0x0010: d4da 94b6 a481 0d02 d802 cc6d 9f85 5106
0x0020: 8c41 d771 4e8f 79cf bf2f 39c9 1dbd ad05
0x0030: 544d 060d 154c 61d2 87d9 e6a2 b17c c353
0x0040: 506c 2b3b e9d5 227f c849 aace 6b8f 3dbc
0x0050: 55fa d232 65dd 51eb 4da5 84fe 95dc bb14
0x0060: 7ba1 4e21 3215 816d c3e9 c7bf 05d9 812b
0x0070: ea8a f5afThe only other thing that jumped out as relevant is that /sys/class/ieee80211/phy0/device/net:wlan0/keys/default/key (also 2/key) had a different value from /sys/class/ieee80211/phy0/sta/00:14:6c:ed:c1:76/key
and the former matched iwconfig's key listing, which had a [3] following it.
Signature · 2 lines
Paul "TBBle" Hampson [email protected]
Please _don't_ remove CCs.
Quoted 12 lines · Paul Hampson
Quoted 9 lines · Michael Buesch <mb <at> bu3sch.de> writes:
Quoted 8 lines · Paul Hampson
I've been backporting the bcm43xx-d80211 driver to whatever the released 2.6 kernel was using the rt2x00 project's d80211 stack (equivalent to current wireless-dev but with a workaround for not having a
ieee80211_devpointer and still using the_tfminterface instead of the_cypherinterface.)As of last night's wireless-dev tree bcm43xx, everything seems to be operating fine except incoming broadcast traffic is coming in 14 bytes too long and scrambled. I presume this means it's not decrypting properly...
Quoted 5 lines
Quoted 2 lines
It sounds like a bug in the hardware decryption setup. Are you using TKIP or not?
Yes, it's using TKIP. The router docs and the loading of the tkip module when I use the softmac driver agree on this.
TKIP is still software encryption. Did you try with WPA-AES, for example, which is hardware encryption?
The problem might be that the card tries to decrypt mcast frames in the crypto hardware, although we did not set a key. So it uses a random key to decrypt. That obviously results in crap.
Signature · 1 line
Greetings Michael.
Quoted 1 line · Michael Buesch
Please _don't_ remove CCs.
Sorry, I was sending via the gmane web interface, I guess it doesn't honour CCs.
This one's via muttng's NNTP support to gmane, so I _think_ it's going to the right places.
Quoted 13 lines · Paul Hampson
Quoted 11 lines
Quoted 9 lines · Michael Buesch <mb <at> bu3sch.de> writes:
Quoted 8 lines · Paul Hampson
I've been backporting the bcm43xx-d80211 driver to whatever the released 2.6 kernel was using the rt2x00 project's d80211 stack (equivalent to current wireless-dev but with a workaround for not having a
ieee80211_devpointer and still using the_tfminterface instead of the_cypherinterface.)As of last night's wireless-dev tree bcm43xx, everything seems to be operating fine except incoming broadcast traffic is coming in 14 bytes too long and scrambled. I presume this means it's not decrypting properly...
Quoted 8 lines
Quoted 5 lines
Quoted 2 lines
It sounds like a bug in the hardware decryption setup. Are you using TKIP or not?
Yes, it's using TKIP. The router docs and the loading of the tkip module when I use the softmac driver agree on this.
TKIP is still software encryption. Did you try with WPA-AES, for example, which is hardware encryption?
The router here only supports TKIP that I can see. There's another network I'll be on on Monday night which I might be able to throw into WPA2 mode... In fact, I was there yesterday and couldn't even get a DHCP IP address, but didn't have the time to diagnose it.
Quoted 3 lines
The problem might be that the card tries to decrypt mcast frames in the crypto hardware, although we did not set a key. So it uses a random key to decrypt. That obviously results in crap.
Eww, yuck.
Signature · 13 lines
----------------------------------------------------------- Paul "TBBle" Hampson, B.Sc, LPI, MCSE On-hiatus Asian Studies student, ANU The Boss, Bubblesworth Pty Ltd (ABN: 51 095 284 361) [email protected] Of course Pacman didn't influence us as kids. If it did, we'd be running around in darkened rooms, popping pills and listening to repetitive music. -- Kristian Wilson, Nintendo, Inc, 1989 License: http://creativecommons.org/licenses/by/2.1/au/ -----------------------------------------------------------
Quoted 28 lines · Paul TBBle Hampson
Quoted 1 line · Michael Buesch
Please _don't_ remove CCs.
Sorry, I was sending via the gmane web interface, I guess it doesn't honour CCs.
This one's via muttng's NNTP support to gmane, so I _think_ it's going to the right places.
Quoted 18 lines · Paul Hampson
Quoted 17 lines
Quoted 6 lines · Michael Buesch <mb <at> bu3sch.de> writes:
Quoted 4 lines · Paul Hampson
I've been backporting the bcm43xx-d80211 driver to whatever the released 2.6 kernel was using the rt2x00 project's d80211 stack (equivalent to current wireless-dev but with a workaround for not having a
ieee80211_devpointer and still using the_tfminterface instead of the_cypherinterface.)Quoted 6 lines
Quoted 3 lines
As of last night's wireless-dev tree bcm43xx, everything seems to be operating fine except incoming broadcast traffic is coming in 14 bytes too long and scrambled. I presume this means it's not decrypting properly...
It sounds like a bug in the hardware decryption setup. Are you using TKIP or not?
Yes, it's using TKIP. The router docs and the loading of the tkip module when I use the softmac driver agree on this.
Quoted 7 lines
Quoted 2 lines
TKIP is still software encryption. Did you try with WPA-AES, for example, which is hardware encryption?
The router here only supports TKIP that I can see. There's another network I'll be on on Monday night which I might be able to throw into WPA2 mode... In fact, I was there yesterday and couldn't even get a DHCP IP address, but didn't have the time to diagnose it.
TKIP hw encryption needs some modifications to the d80211 stack. There are patches available, but I think they are not merged, yet. John, do you know the state on these patches?
Quoted 5 lines
Quoted 3 lines
The problem might be that the card tries to decrypt mcast frames in the crypto hardware, although we did not set a key. So it uses a random key to decrypt. That obviously results in crap.
Eww, yuck.
Ah, and also note that you _need_ firmware from a v4 binary driver to have hardware encryption with bcm43xx.
Signature · 1 line
Greetings Michael.
Hi,
12 Kas 2006 Paz 10:34 tarihinde şunları yazmıştınız:
Quoted 2 lines
Ah, and also note that you _need_ firmware from a v4 binary driver to have hardware encryption with bcm43xx.
Last time I heard only v3 firmware was supported. Is that changed?
Regards,
ismail
Quoted 6 lines · Ismail Donmez
Hi,
12 Kas 2006 Paz 10:34 tarihinde şunları yazmıştınız:Quoted 2 lines
Ah, and also note that you _need_ firmware from a v4 binary driver to have hardware encryption with bcm43xx.
Last time I heard only v3 firmware was supported. Is that changed?
Yes, wireless-dev supports v4 firmware (and even requires it for advanced features like hw encryption)
Signature · 1 line
Greetings Michael.
12 Kas 2006 Paz 23:47 tarihinde, Michael Buesch şunları yazmıştı:
Quoted 11 lines · Ismail Donmez
Quoted 7 lines
Hi,
12 Kas 2006 Paz 10:34 tarihinde şunları yazmıştınız:
Quoted 2 lines
Ah, and also note that you _need_ firmware from a v4 binary driver to have hardware encryption with bcm43xx.
Last time I heard only v3 firmware was supported. Is that changed?
Yes, wireless-dev supports v4 firmware (and even requires it for advanced features like hw encryption)
Thanks for information, this firmware would also solve the recent roothole.
Regards,
ismail
13 Kas 2006 Pts 00:40 tarihinde şunları yazmıştınız:
Quoted 8 lines
12 Kas 2006 Paz 23:47 tarihinde, Michael Buesch şunları yazmıştı:
Quoted 7 lines · Ismail Donmez
Quoted 6 lines
Hi,
12 Kas 2006 Paz 10:34 tarihinde şunları yazmıştınız:
Quoted 2 lines
Ah, and also note that you _need_ firmware from a v4 binary driver to have hardware encryption with bcm43xx.
Quoted 6 lines
Quoted 4 lines
Quoted 1 line
Last time I heard only v3 firmware was supported. Is that changed?
Yes, wireless-dev supports v4 firmware (and even requires it for advanced features like hw encryption)
Thanks for information, this firmware would also solve the recent roothole.
s/roothole/security vulnerability would be better =)
/ismail
Johannes Berg
maintainerQuoted 3 lines · Ismail Donmez
Quoted 1 line
Thanks for information, this firmware would also solve the recent roothole.
s/roothole/security vulnerability would be better =)
Ummm, the problem was in the driver, not the firmware. You can't have security problems like that in the firmware.
johannes
13 Kas 2006 Pts 00:54 tarihinde şunları yazmıştınız:
Quoted 8 lines · Ismail Donmez
Quoted 4 lines
Quoted 2 lines
Thanks for information, this firmware would also solve the recent roothole.
s/roothole/security vulnerability would be better =)
Ummm, the problem was in the driver, not the firmware. You can't have security problems like that in the firmware.
Ah then only ndiswrapper is problematic. Thanks for the insight.
/ismail
Quoted 3 lines · Michael Buesch
TKIP hw encryption needs some modifications to the d80211 stack. There are patches available, but I think they are not merged, yet. John, do you know the state on these patches?
Waiting for Hong Liu to respond to Johannes' comments.
Thanks,
Jiri
Signature · 2 lines
Jiri Benc SUSE Labs
Quoted 15 lines · Michael Buesch
Quoted 14 lines · Paul TBBle Hampson
Quoted 12 lines · Michael Buesch
Quoted 11 lines · Paul Hampson
Quoted 9 lines · Michael Buesch <mb <at> bu3sch.de> writes:
Quoted 8 lines · Paul Hampson
I've been backporting the bcm43xx-d80211 driver to whatever the released 2.6 kernel was using the rt2x00 project's d80211 stack (equivalent to current wireless-dev but with a workaround for not having a
ieee80211_devpointer and still using the_tfminterface instead of the_cypherinterface.)As of last night's wireless-dev tree bcm43xx, everything seems to be operating fine except incoming broadcast traffic is coming in 14 bytes too long and scrambled. I presume this means it's not decrypting properly...
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 5 lines
Quoted 2 lines
It sounds like a bug in the hardware decryption setup. Are you using TKIP or not?
Yes, it's using TKIP. The router docs and the loading of the tkip module when I use the softmac driver agree on this.
TKIP is still software encryption. Did you try with WPA-AES, for example, which is hardware encryption?
I've just tried with WEP-104, which reports hardware encryption for all four keys.
Quoted 2 lines
Ah, and also note that you _need_ firmware from a v4 binary driver to have hardware encryption with bcm43xx.
Just to summarise results so far:
(Current version)
bcm43xx-d80211 w/v3 firmware and TKIP: Garbled broadcast RX bcm43xx-d80211 w/v4 firmware and TKIP: Garbled broadcast RX bcm43xx-d80211 w/v4 firmware and WEP-104: Correct broadcast RX
(version from wireless-dev before October 23rd, unsure of exact date) bcm43xx (softmac) w/v3 firmware and TKIP: Correct broadcast RX
Still to test: bcm43xx-d80211 w/v3 firmware and WEP-104
I also tried looking at how the two drivers I've got here are clearing their keys and comparing them to the bcm specs, and there's nothing obviously wrong that I can see. Sadly I can't compare the drivers directly since one is using the v3 specs and one the v4 specs... -_-
And to top it off, current wireless-dev's softmac bcm43xx driver doesn't build against 2.6.18, and I'm not yet motivated into munging it into doing so, unless it will have diagnostic advantage to this situation. (It'll require also building a new ieee80211softmac at least)
Signature · 13 lines
----------------------------------------------------------- Paul "TBBle" Hampson, B.Sc, LPI, MCSE On-hiatus Asian Studies student, ANU The Boss, Bubblesworth Pty Ltd (ABN: 51 095 284 361) [email protected] Of course Pacman didn't influence us as kids. If it did, we'd be running around in darkened rooms, popping pills and listening to repetitive music. -- Kristian Wilson, Nintendo, Inc, 1989 License: http://creativecommons.org/licenses/by/2.1/au/ -----------------------------------------------------------
Quoted 11 lines · Paul TBBle Hampson
Just to summarise results so far:
(Current version)
bcm43xx-d80211 w/v3 firmware and TKIP: Garbled broadcast RX bcm43xx-d80211 w/v4 firmware and TKIP: Garbled broadcast RX bcm43xx-d80211 w/v4 firmware and WEP-104: Correct broadcast RX(version from wireless-dev before October 23rd, unsure of exact date) bcm43xx (softmac) w/v3 firmware and TKIP: Correct broadcast RX
Still to test: bcm43xx-d80211 w/v3 firmware and WEP-104
bcm43xx-d80211 w/v3 firmware and WEP-104: Garbled _everything_ RX Of course, that's to be expected. WEP only uses broadcast keys, and we knew they were failing in software mode.
Signature · 13 lines
----------------------------------------------------------- Paul "TBBle" Hampson, B.Sc, LPI, MCSE On-hiatus Asian Studies student, ANU The Boss, Bubblesworth Pty Ltd (ABN: 51 095 284 361) [email protected] Of course Pacman didn't influence us as kids. If it did, we'd be running around in darkened rooms, popping pills and listening to repetitive music. -- Kristian Wilson, Nintendo, Inc, 1989 License: http://creativecommons.org/licenses/by/2.1/au/ -----------------------------------------------------------
Johannes Berg
maintainerQuoted 3 lines · Paul TBBle Hampson
bcm43xx-d80211 w/v3 firmware and WEP-104: Garbled _everything_ RX Of course, that's to be expected. WEP only uses broadcast keys, and we knew they were failing in software mode.
Hmm, not sure about this, but it sort of confirms my theory -- with v3 firmware we never upload keys to the hw so the hw will successfully decrypt frames with the 'none' algorithm, and the RX path will treat them as though they were decrypted already.
johannes
Johannes Berg
maintainerQuoted 9 lines · Paul TBBle Hampson
Just to summarise results so far:
(Current version)
bcm43xx-d80211 w/v3 firmware and TKIP: Garbled broadcast RX bcm43xx-d80211 w/v4 firmware and TKIP: Garbled broadcast RX bcm43xx-d80211 w/v4 firmware and WEP-104: Correct broadcast RX(version from wireless-dev before October 23rd, unsure of exact date) bcm43xx (softmac) w/v3 firmware and TKIP: Correct broadcast RX
The last item is interesting. The softmac version doesn't include any hw crypto so it never checks the 'decrypt attempted' bit in the RX header afaik, leaving all crypto to hw.
I postulated before that the problem is that the firmware sets the 'decrypt attempted' bit even if the key is none, hence the driver tells the d80211 stack that the frame was already decrypted (no decrypt error occurs because in reality the algorithm is 'none'.)
You could test this theory by clearing the 'decrypt attempted' bit unconditionally in the RX path before it is ever tested. Then, any wep/aes should no longer work properly with v4 firmware and bcm43xx-d80211, but tkip would. If my theory is correct.
johannes
Quoted 21 lines · Johannes Berg
Quoted 9 lines · Paul TBBle Hampson
Just to summarise results so far:
(Current version)
bcm43xx-d80211 w/v3 firmware and TKIP: Garbled broadcast RX bcm43xx-d80211 w/v4 firmware and TKIP: Garbled broadcast RX bcm43xx-d80211 w/v4 firmware and WEP-104: Correct broadcast RX(version from wireless-dev before October 23rd, unsure of exact date) bcm43xx (softmac) w/v3 firmware and TKIP: Correct broadcast RX
The last item is interesting. The softmac version doesn't include any hw crypto so it never checks the 'decrypt attempted' bit in the RX header afaik, leaving all crypto to hw.
I postulated before that the problem is that the firmware sets the 'decrypt attempted' bit even if the key is none, hence the driver tells the d80211 stack that the frame was already decrypted (no decrypt error occurs because in reality the algorithm is 'none'.)
Quoted 4 lines
You could test this theory by clearing the 'decrypt attempted' bit unconditionally in the RX path before it is ever tested. Then, any wep/aes should no longer work properly with v4 firmware and bcm43xx-d80211, but tkip would. If my theory is correct.
yes, Johannes. The problem is that the decrypt attempted bit is even set for key_none. When I force-disable this codepath, TKIP works perfectly well.
I will do a patch for this. There are a few other minor bugs in the crypto stuff, which I will fix, too. Key clearing is not handled 100% correct, etc...
Signature · 1 line
Greetings Michael.
PATCH 1/1 : Fix bcm43xx-d80211 hwcrypto (was Re: bcm43xx-d80211 broadcast reception with WPA)
In reply to Paul TBBle HampsonPlease test if this fixes hwcrypto.
Note that we will require v4 firmware with this patch. I will remove v3 firmware support, because it's _really_ a huge pain to support both firmware versions.
If someone wants to have v3 support, please fork my tree. Or alternatively simply install v4 firmware, which is a lot easier. ;)
drivers/net/wireless/d80211/bcm43xx/bcm43xx.h +0 -1
@@ -620,7 +620,6 @@ struct bcm43xx_stats {struct bcm43xx_key {u8 enabled; -u8 used;u8 algorithm;u8 address[6 ]; };
drivers/net/wireless/d80211/bcm43xx/bcm43xx_main.c +92 -93
270 lines not on this page · raw message
drivers/net/wireless/d80211/bcm43xx/bcm43xx_xmit.c +33 -23
74 lines not on this page · raw message
Signature · 1 line
Greetings Michael.
Re: [PATCH, RFT]: Fix bcm43xx-d80211 hwcrypto
Quoted 7 lines · Michael Buesch
Please test if this fixes hwcrypto.
Note that we will require v4 firmware with this patch. I will remove v3 firmware support, because it's _really_ a huge pain to support both firmware versions.
If someone wants to have v3 support, please fork my tree. Or alternatively simply install v4 firmware, which is a lot easier. ;)
Tested fine for me with both TKIP sw encryption and WEP-104 hw encryption, on v4 firmware.
(Without applying today's new wireless-dev patches)
Signature · 13 lines
----------------------------------------------------------- Paul "TBBle" Hampson, B.Sc, LPI, MCSE On-hiatus Asian Studies student, ANU The Boss, Bubblesworth Pty Ltd (ABN: 51 095 284 361) [email protected] Of course Pacman didn't influence us as kids. If it did, we'd be running around in darkened rooms, popping pills and listening to repetitive music. -- Kristian Wilson, Nintendo, Inc, 1989 License: http://creativecommons.org/licenses/by/2.1/au/ -----------------------------------------------------------
PATCH 1/1 : bcm43xx-d80211: fix hwcrypto issues (mcast)
In reply to Paul TBBle HampsonThis fixes various bcm43xx-d80211 hwcrypto issues, which mainly prevented mcast frames from being decrypted properly.
This is mostly a rewrite of the key managing code. Note that after this patch v3 firmware is no longer supported.
drivers/net/wireless/d80211/bcm43xx/bcm43xx.h +0 -1
@@ -620,7 +620,6 @@ struct bcm43xx_stats {struct bcm43xx_key {u8 enabled; -u8 used;u8 algorithm;u8 address[6 ]; };
drivers/net/wireless/d80211/bcm43xx/bcm43xx_main.c +92 -93
270 lines not on this page · raw message
drivers/net/wireless/d80211/bcm43xx/bcm43xx_xmit.c +33 -23
74 lines not on this page · raw message
Signature · 1 line
Greetings Michael.
A message that is not in the archive (sent off-list or not yet received): [email protected]
Re: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
Quoted 5 lines · Michael Buesch
This fixes various bcm43xx-d80211 hwcrypto issues, which mainly prevented mcast frames from being decrypted properly.
This is mostly a rewrite of the key managing code. Note that after this patch v3 firmware is no longer supported.
So what is the solution for Apple machines owner who only get a v3 firmware from Apple ? I remember you telling me the answer on irc but I wanted to make it public :-) Some web site we can d/l the windows updates and extract the FW ?
Cheers,
Ben.
Johannes Berg
maintainerRe: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
Quoted 4 lines · Benjamin Herrenschmidt
So what is the solution for Apple machines owner who only get a v3 firmware from Apple ? I remember you telling me the answer on irc but I wanted to make it public :-) Some web site we can d/l the windows updates and extract the FW ?
Yes, fwcutter comes with a huge list of URLs for both firmware versions (I suppose they'll remove the v3 ones now...)
Also, very new Apple drivers (AppleAirportExpress4311 or something like that) for the PCI-E stuff contains v4 firmware, but I doubt you'd get those drivers with an update.
johannes
Re: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
Quoted 9 lines · Johannes Berg
Quoted 4 lines · Benjamin Herrenschmidt
So what is the solution for Apple machines owner who only get a v3 firmware from Apple ? I remember you telling me the answer on irc but I wanted to make it public :-) Some web site we can d/l the windows updates and extract the FW ?
Yes, fwcutter comes with a huge list of URLs for both firmware versions (I suppose they'll remove the v3 ones now...)
Well, the latest "released" version (fwcutter-005) contains a huge list of ... v3 URLs :-) Only 2 v4 in there. I'll check SVN.
Cheers,
Ben.
Re: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
Quoted 10 lines · Benjamin Herrenschmidt
Quoted 9 lines · Johannes Berg
Quoted 4 lines · Benjamin Herrenschmidt
So what is the solution for Apple machines owner who only get a v3 firmware from Apple ? I remember you telling me the answer on irc but I wanted to make it public :-) Some web site we can d/l the windows updates and extract the FW ?
Yes, fwcutter comes with a huge list of URLs for both firmware versions (I suppose they'll remove the v3 ones now...)
v3 ones are still needed for the softmac version, unless I'm badly drunk...
Quoted 2 lines
Well, the latest "released" version (fwcutter-005) contains a huge list of ... v3 URLs :-) Only 2 v4 in there. I'll check SVN.
There's actually a v4 big-endian firmware in the fwcutter list in SVN, but no URL to fetch it from...
It comes from the AppleAirPortBrcm4311 driver. I've no idea where that update comes from, I found a copy on the Internet somewhere. (I think someone posted the file to a message board looking for help with it...)
I'm not sure, would it be be bad for me to just stick a copy of it up on a website? If so, I can prolly dig up the place I got it from (I'm sure I bookmarked it at the time) and people can choose their own level of comfort.
I have to say it's not clear from the documentation that BE machines like PPC need to use a BE firmware... (If that's not the case, it's not clear that it's not the case either, it just doesn't say either way. ^_^)
In fact, given that it's a BE firmware update for Intel-based Macs' bcm cards, I presume the endianess of the firmware doesn't matter?
Signature · 2 lines
Paul Hampson [email protected]
Re: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
In reply to Benjamin HerrenschmidtQuoted 2 lines · Benjamin Herrenschmidt
Well, the latest "released" version (fwcutter-005) contains a huge list of ... v3 URLs :-) Only 2 v4 in there. I'll check SVN.
Still the same. One of them does not exist, the other one requires Javascript!
Andreas.
Signature · 4 lines
Andreas Schwab, SuSE Labs, [email protected] SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany PGP key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5 "And now for something completely different."
A message that is not in the archive (sent off-list or not yet received): [email protected]
Re: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
Quoted 2 lines · Andreas Schwab
Still the same. One of them does not exist, the other one requires Javascript!
Yeah, looks like Martin forgot to put the URLs in the the readme while he put them into commit messages...
Try these:
Support for bcmwl5.sys v4.80.53.0 added. ftp://downloads.netgear.com/files/wn511b_sw_3_28_3_8_setup.zip (bcmwl5.sys is renamed to wn511b.sys)
Support for bcmwl5(64).sys v4.100.15.5 added. from the latest Linksys WPC300N driver at
http://www.linksys.com/download/
It's ucode revision 0x0173, patchlevel 0x0425, 2006-10-04.
Martin, can you add these to the readme?
johannes
Re: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
Quoted 5 lines · Johannes Berg
Try these:
Support for bcmwl5.sys v4.80.53.0 added. ftp://downloads.netgear.com/files/wn511b_sw_3_28_3_8_setup.zip (bcmwl5.sys is renamed to wn511b.sys)
No supported files in there.
Quoted 4 lines
Support for bcmwl5(64).sys v4.100.15.5 added. from the latest Linksys WPC300N driver at
http://www.linksys.com/download/
It's ucode revision0x0173, patchlevel0x0425, 2006-10-04.
Requires Javascript.
Andreas.
Signature · 4 lines
Andreas Schwab, SuSE Labs, [email protected] SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany PGP key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5 "And now for something completely different."
Re: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
In reply to Johannes BergQuoted 7 lines · Johannes Berg
Quoted 2 lines · Andreas Schwab
Still the same. One of them does not exist, the other one requires Javascript!
Yeah, looks like Martin forgot to put the URLs in the the readme while he put them into commit messages...
The last update for README was quite some time ago. I think that file is a bad way to transport that kind of information. So I'm afraid a lot of links are really out of date now. A better way is probably an open wiki (berlios?). Volunteers and other ideas are always welcome.
The comments in the commit messages are more a kind of protocol and not the best link on earth. And if you wait some days the number of mirrors for the latest driver will explode. It would be bad to write the first available link into the README. A download link should be a direct link, a small package, easy to extract and so on. But I'm really not interested in doing that job of a hunter for the perfect download link and try to keep that link up to date. I think it's more important to write support for more driver files instead of doing that surrounding stuff for a few files in a perfect way. Google isn't that complicated these days and it's more up to date than our README.
Quoted 10 lines
Support for bcmwl5.sys v4.80.53.0 added. ftp://downloads.netgear.com/files/wn511b_sw_3_28_3_8_setup.zip (bcmwl5.sys is renamed to wn511b.sys)
Support for bcmwl5(64).sys v4.100.15.5 added.
Quoted 1 line
from the latest Linksys WPC300N driver at
http://www.linksys.com/download/
It's ucode revision0x0173, patchlevel0x0425, 2006-10-04.Martin, can you add these to the readme?
These are bad examples for adding to README. People aren't prepared well enough to look for renamed bcmwl5.sys files which are packaged twice and they expect direct links, too. ...
Martin
Re: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
In reply to Andreas SchwabQuoted 7 lines · Andreas Schwab
Quoted 2 lines · Benjamin Herrenschmidt
Well, the latest "released" version (fwcutter-005) contains a huge list of ... v3 URLs :-) Only 2 v4 in there. I'll check SVN.
Still the same. One of them does not exist, the other one requires Javascript!
MacOS X 10.4.8 Update for Intel [1] has AppleAirportBrcm4311 E99AD02FDD7699FA52F0B9153C8411C8
4.80.46.0 big-endian, supported in fwcutter-svn
It's a 200MB download for just the one file, but w3m seems to be able to process it OK, so I assume it doesn't require Javascript.
[1] http://www.apple.com/support/downloads/macosx1048updateintel.html
Signature · 13 lines
----------------------------------------------------------- Paul "TBBle" Hampson, B.Sc, LPI, MCSE On-hiatus Asian Studies student, ANU The Boss, Bubblesworth Pty Ltd (ABN: 51 095 284 361) [email protected] Of course Pacman didn't influence us as kids. If it did, we'd be running around in darkened rooms, popping pills and listening to repetitive music. -- Kristian Wilson, Nintendo, Inc, 1989 License: http://creativecommons.org/licenses/by/2.1/au/ -----------------------------------------------------------
Re: [PATCH]: bcm43xx-d80211: fix hwcrypto issues (mcast)
In reply to Benjamin HerrenschmidtQuoted 11 lines · Benjamin Herrenschmidt
Quoted 5 lines · Michael Buesch
This fixes various bcm43xx-d80211 hwcrypto issues, which mainly prevented mcast frames from being decrypted properly.
This is mostly a rewrite of the key managing code. Note that after this patch v3 firmware is no longer supported.
So what is the solution for Apple machines owner who only get a v3 firmware from Apple ? I remember you telling me the answer on irc but I wanted to make it public :-) Some web site we can d/l the windows updates and extract the FW ?
http://bcm43xx.spugna.org/index.php?topic=141.msg517 provides the following links to 4.80.46.0:
http://www.i-nz.net/files/perm/AppleAirPortBrcm4311.tar.gz
http://www.i-nz.net/files/perm/bcmwl5-win32.tar.gz
Signature · 13 lines
----------------------------------------------------------- Paul "TBBle" Hampson, B.Sc, LPI, MCSE On-hiatus Asian Studies student, ANU The Boss, Bubblesworth Pty Ltd (ABN: 51 095 284 361) [email protected] Of course Pacman didn't influence us as kids. If it did, we'd be running around in darkened rooms, popping pills and listening to repetitive music. -- Kristian Wilson, Nintendo, Inc, 1989 License: http://creativecommons.org/licenses/by/2.1/au/ -----------------------------------------------------------