Return-path: Received: from ti-out-0910.google.com ([209.85.142.186]:8699 "EHLO ti-out-0910.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753165AbYKNSRG (ORCPT ); Fri, 14 Nov 2008 13:17:06 -0500 Received: by ti-out-0910.google.com with SMTP id b6so920161tic.23 for ; Fri, 14 Nov 2008 10:17:04 -0800 (PST) From: Sujith MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Message-ID: <18717.49111.572993.44410@localhost.localdomain> (sfid-20081114_191711_705877_0B70AEB0) Date: Fri, 14 Nov 2008 23:43:43 +0530 To: "Luis R. Rodriguez" Cc: "Bob Copeland" , "Dan McGee" , linux-wireless@vger.kernel.org, "Michael Buesch" , "Johannes Berg" Subject: Re: Kernel oops when loading ath5k from compat-wireless in 2.6.27 In-Reply-To: <43e72e890811141006x1c396c78ube87a526f1435676@mail.gmail.com> References: <449c10960811132146s40aef6c6ue8dfeef5ba29812a@mail.gmail.com> <43e72e890811132217k160db63ch77e7d03c38e81d5f@mail.gmail.com> <43e72e890811140937h4dc03cf8g50df7f74efcd28b2@mail.gmail.com> <43e72e890811141006x1c396c78ube87a526f1435676@mail.gmail.com> Sender: linux-wireless-owner@vger.kernel.org List-ID: Luis R. Rodriguez wrote: > On Fri, Nov 14, 2008 at 9:49 AM, Bob Copeland wrote: > > On Fri, Nov 14, 2008 at 12:37 PM, Luis R. Rodriguez wrote: > >> Anyway we do set the parent before calling ieee80211_register_hw() by > >> using SET_IEEE80211_DEV(hw, &pdev->dev) (oh ok here is the parent). so > >> when we try to get the name with parent->driver I am not sure if > >> driver has been set yet because ath5k_pci_probe() hasn't finished yet > >> as probe called ath5k_hw_attach(). When does ->driver get set and why > >> would it fail only for ath5k? > > > > Ahh, I believe that would be in pci-driver.c __pci_device_probe, after driver > > probe is run successfully. Yes, we do ieee80211_register_hw from within > > probe... So, how did this ever work again? /me scratches head. > > Wow.. ok, yeah I see that, this is a mac8021 bug then unless mac80211 > wants us to not callieee80211_register_hw() from probe or its children > which would be weird. >From the original log: ath5k_pci 0000:01:00.0: registered as '' Sujith