Re: [BUG]: 3.7.0.rc2 iwlwifi: Mn=onitor mode for Intel 6235 includes FCS in frames

Re: [BUG]: 3.7.0.rc2 iwlwifi: Mn=onitor mode for Intel 6235 includes FCS in frames

[Johannes Berg]

Hi Mark, sorry for the delay,

On Mon, 2012-10-29 at 11:56 +0000, MARK PHILLIPS wrote:
> I have found bug which breaks monitor mode for the Intel Centrino
> Advanced-N 6235. When in monitor mode the FCS is returned appended to
> the frames which confuses wireshark.
> 
> The fix appears to be trivial, just add IEEE80211_HW_RX_INCLUDES_FCS
> to the hw->flags when in monitor mode via a change to
> drivers/net/wireless/iwlwifi/dvm/mac80211.c

Yes, this seems to be a reasonable change, unfortunately mac80211
doesn't provide a per packet flag whether the FCS is present or not.

I think I prefer doing the modification in iwlagn_mac_add_interface(),
and will also add a comment.

FWIW, this appears to be a regression introduced in
578977264199de9815ace51ade87cec4894cf010, which went into 3.5. I'm a bit
confused why this didn't appear in my tests though, so I'm trying to see
if different firmware versions behave differently.

johannes

RE: [BUG]: 3.7.0.rc2 iwlwifi: Mn=onitor mode for Intel 6235 includes FCS in frames

[Mark Phillips]

Great, glad you are sorting it out, thanks for all your help!

Cheers
Mark

-----Original Message-----
From: Johannes Berg [mailto:johannes@sipsolutions.net] 
Sent: 13 November 2012 20:29
To: MARK PHILLIPS
Cc: linux-wireless@vger.kernel.org
Subject: Re: [BUG]: 3.7.0.rc2 iwlwifi: Mn=onitor mode for Intel 6235 includes FCS in frames

Hi Mark, sorry for the delay,

On Mon, 2012-10-29 at 11:56 +0000, MARK PHILLIPS wrote:
> I have found bug which breaks monitor mode for the Intel Centrino 
> Advanced-N 6235. When in monitor mode the FCS is returned appended to 
> the frames which confuses wireshark.
> 
> The fix appears to be trivial, just add IEEE80211_HW_RX_INCLUDES_FCS 
> to the hw->flags when in monitor mode via a change to 
> drivers/net/wireless/iwlwifi/dvm/mac80211.c

Yes, this seems to be a reasonable change, unfortunately mac80211 doesn't provide a per packet flag whether the FCS is present or not.

I think I prefer doing the modification in iwlagn_mac_add_interface(), and will also add a comment.

FWIW, this appears to be a regression introduced in 578977264199de9815ace51ade87cec4894cf010, which went into 3.5. I'm a bit confused why this didn't appear in my tests though, so I'm trying to see if different firmware versions behave differently.

johannes