From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail.candelatech.com ([208.74.158.172]:33950 "EHLO ns3.lanforge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752591AbaAHSgz (ORCPT ); Wed, 8 Jan 2014 13:36:55 -0500 Message-ID: <52CD9AAF.1040407@candelatech.com> (sfid-20140108_193658_955410_CA0D38CD) Date: Wed, 08 Jan 2014 10:36:31 -0800 From: Ben Greear MIME-Version: 1.0 To: "linux-wireless@vger.kernel.org" CC: ath10k@lists.infradead.org Subject: How to make software re-calc received frames with bad FCS? Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-wireless-owner@vger.kernel.org List-ID: I'm back to trying to get multiple stations working to same AP with WPA2 on ath10k. I tried doing 'nohwcrypt', but I had no luck with that. I think the firmware would require a fair bit of changes to make that happen. So, a second plan is to just tell the firmware to use hardware crypt for the VIFs. This would have a good benefit of offloading the tx path, at least. This seems to work for transmit, but on receive, the hardware hash collides since 2+ stations have same peer (ie, the AP). I do receive packets, but they are flagged with FCS errors (and I think there is no re-assembly of ampdu, but I'm not sure how much that matters). My thought is to somehow flag these packets as needing to be decrypted by software instead, similar to how the normal nohwcrypt logic works. Any suggestions for how to go about attempting this? Maybe just force remove the CRC failure bit from the status before passing up the stack? Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com