On Fri, Jun 27, 2008 at 9:14 PM, Michael Buesch <[EMAIL PROTECTED]> wrote: > On Friday 27 June 2008 19:19:41 Stefanik Gábor wrote: >> On Tue, May 20, 2008 at 6:24 PM, Stefanik Gábor <[EMAIL PROTECTED]> wrote: >> > Hi all, >> > >> > One day I was testing monitor-while-operating, with "wlan0" as the >> > managed interface and "rtap0" as the monitor one, and found an >> > interesting bug: when I am associated with an AP on wlan0, and try to >> > receive the same AP's packets through the monitor interface, the >> > packets arrive in decrypted form, but with the WEP bit still set. This >> > appears to be a HW-crypto-related bug, though I haven't tested with HW >> > crypto off. What can cause such a bug? >> > >> > Thanks, >> > Gábor >> > >> > -- >> > Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-) >> > >> >> Could someone investigate this? I first posted this a month ago, and >> noone really did anything about it. > > I'm sorry, we cannot reproduce this bug. > I'm also unsure how hwcrypto could break here. The driver/firmware doesn't > even know about whether there's a monitor interface in addition to the > STA interface. The firmware simply decrypts the packet or it doesn't. > If the firmware does decrypt, the driver will clear the protected bit. > If it doesn't decrypt, the driver won't clear the bit. > > I guess you need to debug this by yourself, if you can still reproduce > the bug. see drivers/net/b43/xmit.c for the b43_rx function. You can add > a few sanity checks there. If you don't find a bug here, go up to the > mac80211 stack. > > -- > Greetings Michael. >
Hmm... I re-checked, and now it appears that the packets are now passed with the protected bit set correctly, but still they arrive decrypted. Trying to transmit an already-encrypted frame results in it being double-encrypted. However, I believe that a raw monitor mode interface should pass the packets as they are received. and not in decrypted form. -- Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-) _______________________________________________ Bcm43xx-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
