From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mx1.redhat.com ([209.132.183.28]:51981 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751016Ab3DOOht (ORCPT ); Mon, 15 Apr 2013 10:37:49 -0400 Message-ID: <1366036772.2404.3.camel@dcbw.foobar.com> (sfid-20130415_163753_659716_F0EBFC0E) Subject: Re: Centrino Ultimate-N 6300 AGN rfkill always Soft blocked From: Dan Williams To: Johannes Berg Cc: schwaahed , linux-wireless@vger.kernel.org Date: Mon, 15 Apr 2013 09:39:32 -0500 In-Reply-To: <1366032132.8361.19.camel@jlt4.sipsolutions.net> References: <1365804259.10342.20.camel@dcbw.foobar.com> (sfid-20130413_000245_000077_30E72639) <1366032132.8361.19.camel@jlt4.sipsolutions.net> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, 2013-04-15 at 15:22 +0200, Johannes Berg wrote: > On Fri, 2013-04-12 at 17:04 -0500, Dan Williams wrote: > > On Thu, 2013-04-11 at 20:03 -0400, schwaahed wrote: > > > Hello, > > > > > > I have an HP Elitebook 8460w laptop containing a Centrino Ultimate-N > > > 6300 AGN (linux-3.8.6 iwlwifi module - firmware version 9.221.4.1 > > > build 25532) > > > > I notice that your 'rfkill list all' dumps below don't contain an entry > > for hp-wmi. Is that module loaded on your machine? It's needed at > > least for earlier models of the Elitebook, up at least until the xx40 > > series. Perhaps it needs to be updated for your laptop? You might ask > > on the general kernel lists if anyone knows how rfkill is supposed to > > work for the 8460 and if hp-wmi supports it, or if some other mechanism > > is used. > > Yeah it seems not loaded, but how then does the soft block affect hard > block? All very odd, but almost certainly a platform rather than a wifi > issue. On my 2530p and a lot of other laptops, blocking the BIOS "switch" (either physically or softblock) hardblocks phy0. I guess the BIOS twiddles a GPIO that's connected to the mPCI-E module's rfkill line/GPIO? Which leads to the problem we've talked about a long time ago; you can't treat phy0 hardblock as a physical block that cannot be soft-unblocked, because some other switch might actually control it's state. We tried to gray-out the "Enable Wireless" when hardblocked (since logically you can't soft-unblock something that's hardblocked), but it turns out you can't do that because unblocking BIOS switches might un-hardblock the phy0 wifi switch... (and the kernel doesn't describe these dependencies, because, well, that's laptop-specific and would never be up-to-date). Dan