From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S267759AbUBTJ1K (ORCPT ); Fri, 20 Feb 2004 04:27:10 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S267758AbUBTJ1K (ORCPT ); Fri, 20 Feb 2004 04:27:10 -0500 Received: from caramon.arm.linux.org.uk ([212.18.232.186]:22545 "EHLO caramon.arm.linux.org.uk") by vger.kernel.org with ESMTP id S267759AbUBTJ1H (ORCPT ); Fri, 20 Feb 2004 04:27:07 -0500 Date: Fri, 20 Feb 2004 09:26:51 +0000 From: Russell King To: David Brownell Cc: Benjamin Herrenschmidt , Linus Torvalds , Greg KH , Andrew Morton , Linux-USB , Kernel Mailing List Subject: Re: [linux-usb-devel] Re: [BK PATCH] USB update for 2.6.3 Message-ID: <20040220092651.B22235@flint.arm.linux.org.uk> Mail-Followup-To: David Brownell , Benjamin Herrenschmidt , Linus Torvalds , Greg KH , Andrew Morton , Linux-USB , Kernel Mailing List References: <20040220012802.GA16523@kroah.com> <1077256996.20789.1091.camel@gaston> <1077258504.20781.1121.camel@gaston> <4035C068.8050605@pacbell.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <4035C068.8050605@pacbell.net>; from david-b@pacbell.net on Fri, Feb 20, 2004 at 12:08:08AM -0800 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Feb 20, 2004 at 12:08:08AM -0800, David Brownell wrote: > Benjamin Herrenschmidt wrote: > > Yes. I also remember a time where the dma mask for the DMA API was all > > broken too (would not be possible to map the PCI one on top of it), > > but I think that got fixed. > > I thought it was still broken. Last I compared different asm-* arch > implementations of dma_supported(), the semantics were inconsistent. > The "mask" was sometimes ignored, sometimes treated as an upper > bound on the address, once I recall it even being used as a mask! > Some of that inconsistency seemed to come from PCI though. My understanding is that it's a mask, which just happens to be treated as an upper address limit on sane platforms. However, we have a case where it's a real mask (no surprises there!) Intel decided not to fix a bug in a chip which means that an address line (normally A20) must never be asserted during DMA cycles, although it can address more than 1MB of memory. The effect of this is that even MB are dma-able and odd MB aren't. Not nice. -- Russell King Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/ maintainer of: 2.6 PCMCIA - http://pcmcia.arm.linux.org.uk/ 2.6 Serial core