Return-path: Received: from s3.sipsolutions.net ([5.9.151.49]:51891 "EHLO sipsolutions.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752998AbbFIHao (ORCPT ); Tue, 9 Jun 2015 03:30:44 -0400 Message-ID: <1433835041.1892.3.camel@sipsolutions.net> (sfid-20150609_093048_203667_AE9891DB) Subject: Re: [PATCH] Add new mac80211 driver mwlwifi. From: Johannes Berg To: David Lin Cc: "linux-wireless@vger.kernel.org" , Pete Hsieh , Chor Teck Law Date: Tue, 09 Jun 2015 09:30:41 +0200 In-Reply-To: <52abf3860d5941869a72b4ecf9c99616@SC-EXCH02.marvell.com> References: <6d9f23a4a8bf4315b0c7630a9a1c37ad@SC-EXCH02.marvell.com> (sfid-20150604_065726_449748_FA29E81C) <1433598191.2467.65.camel@sipsolutions.net> <52abf3860d5941869a72b4ecf9c99616@SC-EXCH02.marvell.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Tue, 2015-06-09 at 07:25 +0000, David Lin wrote: > 1. That's interesting, why does a PCI(e) driver need OF? > This driver will accept parameters in DTS file for band control, > antenna setting and power table, to cater for different boards > combination inside a system. Ok. I saw some OF references at the very end, but then got a bit side-tracked :) > 2. Can you perhaps explain how the STA/AP firmware separation works? > For this driver, the AP/STA mode will use the same single firmware > binary, so there is no “separation” per se. The firmware will support > AP/STA mode. Hmm, interesting. It seemed that there were references in the code to loading different firmware? Might want to look through for that. > 3. Does this driver has any relation to mwifiex? > Mwifiex is driver for Firmware-based MLME. It interfaces with firmware > with 802.3 packets. Mwlwifi is the Host Soft AP/STA driver that works > with mac80211. Ok, so basically no relation - makes sense. > 1. Directory name “mwlwifi”. > This is to be consistent with some predecessor. For example, Intel > uses iwlwifi, realtek uses rtlwifi. We do not see a need to change it. > This will make sure we keep the original project mwlwifi on openwrt > folder remain intact, while we continue to maintain them the same > way/pace. That's fine. > 2. Interface with F/W. > F/W used by this driver is also used by other marvell’s drivers. Yeah, but I'm not saying you should change the interface, I'm just saying you need to properly endian-annotate it. That won't actually change the binary interface, but will make the driver more robust. > 3. AMPDU related code. > It has been well tested and leveraged from mwl8k. We may enhance it in > future, but please accept the current code status for now. I'm not really sure it's correct, but I guess I can't really comment too much on it. johannes