Return-path: Received: from mail-pz0-f46.google.com ([209.85.210.46]:57290 "EHLO mail-pz0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752658Ab0GZQ4T convert rfc822-to-8bit (ORCPT ); Mon, 26 Jul 2010 12:56:19 -0400 MIME-Version: 1.0 In-Reply-To: References: <4C3D5F3F.7070901@odi.ch> <4C499E6F.8030507@odi.ch> From: "Luis R. Rodriguez" Date: Mon, 26 Jul 2010 09:55:58 -0700 Message-ID: Subject: Re: BUG at arch/x86/mm/physaddr.c:5 To: Jike Song Cc: =?UTF-8?Q?Ortwin_Gl=C3=BCck?= , linux-kernel@vger.kernel.org, Felix Fietkau , Vasanthakumar Thiagarajan , jmalinen@atheros.com, Sujith.Manoharan@atheros.com, senthilkumar@atheros.com, linux-wireless@vger.kernel.org Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: 2010/7/25 Jike Song : > 2010/7/23 Ortwin Glück : >> On 23.07.2010 10:04, Jike Song wrote: >>> Seems to be a logical error?  Does the following patch remove your Oops? >> >> Looking at the init code I agree. Yes, the patch fixes the OOPS. I am sure >> because the OOPS occurred at every boot and now it's gone. >> >> Thanks! >> >> Ortwin > > Hi Maintainers, > >  Since Ortwin has confirmed this patch , would you give me an ACK/NAK please? > > > -- > Thanks, > Jike > > > From 1414829dcc5b53ef4802f5de2b9be58e2ca0fb23 Mon Sep 17 00:00:00 2001 > From: Jike Song > Date: Mon, 26 Jul 2010 13:42:41 +0800 > Subject: [PATCH] ath9k: fix wrong DMA direction in RX tasklet > MIME-Version: 1.0 > Content-Type: text/plain; charset=utf-8 > Content-Transfer-Encoding: 8bit > > Commit b5c80475abaad015699384ca64ef8229fdd88758 introduced edma > support for RX, hence bidirectional DMA support. But it > specified the DMA type as DMA_FROM_DEVICE by mistake. > > Reported-and-tested-by: Ortwin Glück > Signed-off-by: Jike Song > --- >  drivers/net/wireless/ath/ath9k/recv.c |    4 ++-- >  1 files changed, 2 insertions(+), 2 deletions(-) > > diff --git a/drivers/net/wireless/ath/ath9k/recv.c > b/drivers/net/wireless/ath/ath9k/recv.c > index ca6065b..e3e5291 100644 > --- a/drivers/net/wireless/ath/ath9k/recv.c > +++ b/drivers/net/wireless/ath/ath9k/recv.c > @@ -844,9 +844,9 @@ int ath_rx_tasklet(struct ath_softc *sc, int flush, bool hp) >        int dma_type; > >        if (edma) > -               dma_type = DMA_FROM_DEVICE; > -       else >                dma_type = DMA_BIDIRECTIONAL; > +       else > +               dma_type = DMA_FROM_DEVICE; > >        qtype = hp ? ATH9K_RX_QUEUE_HP : ATH9K_RX_QUEUE_LP; >        spin_lock_bh(&sc->rx.rxbuflock); > -- I'll let Felix chime in on this one as he last reviewed this path. There were some serious issues with MIPs and DMA that at this point I think only he grocked. Luis