From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from p508B7B47.dip.t-dialin.net ([80.139.123.71]:3621 "EHLO mail.linux-mips.net") by vger.kernel.org with ESMTP id S261202AbUCPLSC (ORCPT ); Tue, 16 Mar 2004 06:18:02 -0500 Date: Tue, 16 Mar 2004 12:17:47 +0100 From: Ralf Baechle Subject: Re: More DMA API junk Message-ID: <20040316111747.GA1592@linux-mips.org> References: <1079398383.1968.205.camel@gaston> <1079399555.1804.232.camel@mulgrave> <1079400742.1968.209.camel@gaston> <1079407473.2069.383.camel@mulgrave> <1079407959.2348.226.camel@gaston> <1079408930.2158.408.camel@mulgrave> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1079408930.2158.408.camel@mulgrave> To: James Bottomley Cc: Benjamin Herrenschmidt , Linux Arch list List-ID: On Mon, Mar 15, 2004 at 10:48:48PM -0500, James Bottomley wrote: > You're confused about what it's used for. It's designed only to be > called on memory allocated by dma_alloc_noncoherent() and tells you if > that API actually returned coherent memory or not. This was designed ... but what is coherent? The API's function naming is completly confusing at least from a MIPS perspective. MIPS calls something coherent if cache coherency is maintained by hardware and that's (hey, what caches?!?) not the case for uncached memory. So I would like to rename functions to something less confusing. Where did the original terminology come from? Would anybody object simply swapping the coherent and noncoherent parts of the function names? Ralf