From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail-ee0-f48.google.com ([74.125.83.48]:52251 "EHLO mail-ee0-f48.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753684AbaBLU2E (ORCPT ); Wed, 12 Feb 2014 15:28:04 -0500 Received: by mail-ee0-f48.google.com with SMTP id t10so4557821eei.7 for ; Wed, 12 Feb 2014 12:28:02 -0800 (PST) Message-ID: <52FBD950.7080204@gmail.com> (sfid-20140212_212808_654604_93E4D9CF) Date: Wed, 12 Feb 2014 22:28:00 +0200 From: Emmanuel Grumbach MIME-Version: 1.0 To: Stanislaw Gruszka , "Grumbach, Emmanuel" CC: "linux-wireless@vger.kernel.org" , "Intel Linux Wireless (ilw@linux.intel.com)" Subject: Re: [RFC] iwlwifi: refactor the TX / RX ampdu override References: <1392196175-6219-1-git-send-email-egrumbach@gmail.com> <1392203989-15645-1-git-send-email-emmanuel.grumbach@intel.com> <20140212121314.GE1315@redhat.com> <0BA3FCBA62E2DC44AF3030971E174FB303D66597@HASMSX103.ger.corp.intel.com> <20140212140420.GA3379@redhat.com> In-Reply-To: <20140212140420.GA3379@redhat.com> Content-Type: text/plain; charset=us-ascii Sender: linux-wireless-owner@vger.kernel.org List-ID: On 02/12/2014 04:04 PM, Stanislaw Gruszka wrote: > On Wed, Feb 12, 2014 at 12:27:05PM +0000, Grumbach, Emmanuel wrote: >>> I think we can keep IWL_DISABLE_HT_ALL option. I should be possible to >>> make some per-device type default settings and if disable_11n == 0 use >>> them. Otherwise use settings from module parameter. Will that work ? >> >> I'd prefer to have the default based on the iwldvm vs. iwlmvm instead of adding yet another per-HW field. But yeah - it is possible. > > I re-think changing default settings do disable TX AGG and must tell > that this is very odd. I would prefer to stay defaults as they are now. > Users can disable TX aggregation using module option. I disagree. We have bugs there - it is pretty much obvious. I remember a tracing from 2.6.39 which was before the re-architecture I made in the driver with clear FW bugs. I don't remember what device though. This is why I want to disable this by default. You asked me to do this with a module parameter and not with a Kconfig option - I agreed. Now - if you want to ship with a different settings, you can... but we both know the risk. > > I also not sure if that is really a firmware bug, it is rather more > probable that TX aggregation worked pretty fine on 5100 and other older > devices, but some driver changes broke that. Frankly, I don't really remember that... I have to say that I might not have been in the business at that period...