From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Morton Subject: Re: [PATCH 05/37] ata: use get/put_endian helpers Date: Thu, 29 May 2008 21:08:59 -0700 Message-ID: <20080529210859.0feee69d.akpm@linux-foundation.org> References: <1212092282.28403.107.camel@brick> <483F65CA.4050507@rtr.ca> <20080529195225.5665ae9c.akpm@linux-foundation.org> <1212118661.28403.175.camel@brick> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1212118661.28403.175.camel@brick> Sender: linux-ide-owner@vger.kernel.org To: Harvey Harrison Cc: Mark Lord , linux-arch , Jeff Garzik , linux-ide List-Id: linux-arch.vger.kernel.org On Thu, 29 May 2008 20:37:40 -0700 Harvey Harrison wrote: > On Thu, 2008-05-29 at 19:52 -0700, Andrew Morton wrote: > > On Thu, 29 May 2008 22:26:18 -0400 Mark Lord wrote: > > > > > otoh, > > > > - I think the args are backwards > > > > It was made to look like put_unaligned_* and put_unaligned() > > I think of it as put_le16(source, dest) fn(dest, src) is a well-known C idiom which the kernel went and broke in lots of places :( It maps nicely onto dest = src as a mnemonic. But given that the compiler will reliably barf if they are backwards it's a minor thing. > > - I don't like the use of the put_*() namespace. It makes it look > > like a uaccess operation. > > I'm flexible...although I'd probably get nailed to a cross for changing > it again so soon. (the unaligned versions, that is) We'll live. (I'm not a big fan of the patches, btw. It's more of an "oh groan, because it's there I suppose we really ought to do this" sort of thing. (strong objections would save me a lot of work (looks hopefully at inbox))). From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp1.linux-foundation.org ([140.211.169.13]:45260 "EHLO smtp1.linux-foundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758770AbYE3EJh (ORCPT ); Fri, 30 May 2008 00:09:37 -0400 Date: Thu, 29 May 2008 21:08:59 -0700 From: Andrew Morton Subject: Re: [PATCH 05/37] ata: use get/put_endian helpers Message-ID: <20080529210859.0feee69d.akpm@linux-foundation.org> In-Reply-To: <1212118661.28403.175.camel@brick> References: <1212092282.28403.107.camel@brick> <483F65CA.4050507@rtr.ca> <20080529195225.5665ae9c.akpm@linux-foundation.org> <1212118661.28403.175.camel@brick> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-arch-owner@vger.kernel.org List-ID: To: Harvey Harrison Cc: Mark Lord , linux-arch , Jeff Garzik , linux-ide Message-ID: <20080530040859.g0girEG3oJKyzTeJrq4TzUC84nUad7kyhbAXh2z5hYw@z> On Thu, 29 May 2008 20:37:40 -0700 Harvey Harrison wrote: > On Thu, 2008-05-29 at 19:52 -0700, Andrew Morton wrote: > > On Thu, 29 May 2008 22:26:18 -0400 Mark Lord wrote: > > > > > otoh, > > > > - I think the args are backwards > > > > It was made to look like put_unaligned_* and put_unaligned() > > I think of it as put_le16(source, dest) fn(dest, src) is a well-known C idiom which the kernel went and broke in lots of places :( It maps nicely onto dest = src as a mnemonic. But given that the compiler will reliably barf if they are backwards it's a minor thing. > > - I don't like the use of the put_*() namespace. It makes it look > > like a uaccess operation. > > I'm flexible...although I'd probably get nailed to a cross for changing > it again so soon. (the unaligned versions, that is) We'll live. (I'm not a big fan of the patches, btw. It's more of an "oh groan, because it's there I suppose we really ought to do this" sort of thing. (strong objections would save me a lot of work (looks hopefully at inbox))).