[ath9k-devel] Unusal question: ATH9k and broken-on-purpose 802.11 frames

[ath9k-devel] Unusal question: ATH9k and broken-on-purpose 802.11 frames

[jpo]

Hello all,
due to difficult to explain reasons that involve customers that believe in security by obscurity I made some modifications to the then current Madwifi driver for older Atheros chips. Specifically, I changed some values in the frame control field (version number, beacon and probe response values). This way IBSS networks with our modifications would not show up in a normal scan.

Fast forward to 2016: It's becoming harder to source the old hardware (both the WLAN card and actual CPU boards that support miniPCI instead of PCIe). That's why we have to upgrade our hardware. Selecting a WLAN card seems to be the most troublesome part, because most cards only support consumer grade instead of the extended industrial specs (-40..85 ?C) we need. We have narrowed the field down to ath9k or ath10k based cards (BTW: I'm open to alternative suggestions here!). ATH10K is problematic, because it would need a modified firmware blob and doesn't support IBSS out of the box.

That's why our strong preference is to go with ath9k. Since the very first Atheros 802.11n cards were supported by the old Madwifi driver, I would assume that there are no major obstacles to support the broken-on-purpose 802.11 frames. Am I right? Is there anything that I should keep in mind? Does the hardware filter invalid frames? If so, can I disable or change these filters?

Thanks in advance and kind regards
  Joerg