From mboxrd@z Thu Jan 1 00:00:00 1970 From: Larry Finger Subject: Re: [OOPS] bcm43xx oops on 2.6.20-rc1 on x86_64 Date: Sat, 30 Dec 2006 23:14:58 -0600 Message-ID: <45974752.90601@lwfinger.net> References: <20061230192104.GB20714@stusta.de> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Adrian Bunk , linux-kernel@vger.kernel.org, st3@riseup.net, linville@tuxdriver.com, netdev@vger.kernel.org, jgarzik@pobox.com Return-path: To: Aaron Sethman In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Aaron Sethman wrote: > > On Sat, 30 Dec 2006, Adrian Bunk wrote: > >> On Sun, Dec 17, 2006 at 03:15:28PM -0500, Aaron Sethman wrote: >>> >>> Just was loading the bcm43xx module and got the following oops. Note >>> that >>> this card is one of the newer PCI-E cards. If any other info is needed >>> let me know. >> >> Is this issue still present in 2.6.10-rc2-git1? >> >> If yes, was 2.6.19 working fine? >> > > This seems to be fixed in 2.6.20-rc2-git1. Still having other issues > with the driver, but the oops in the SoftMAC code is resolved now at least. I have just started working with the PCI-E BCM4311 that is in my new computer. It receives data OK, but there is something wrong with the DMA out stuff in bcm43xx-softmac - at least for x86_64. All the slots get full but nothing is ever transmitted. FWIW, the wireless-dev git tree works. I'm using it for communications now. I'm using openSUSE 10.2 which uses NetworkManager to configure my WPA-PSK TKIP encrypted network. The signal strengths are roughly the same as I got for my old BCM4306 card. I will post any fixes for the DMA problem as soon as they are available, but it may be a while. I will be off-line until Thursday while I attend a funeral. Larry