From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?ISO-8859-1?Q?Richard_R=F6jfors?= Subject: Re: [PATCH 2/4] xilinx_spi: Switch to iomem functions and support little endian. Date: Wed, 11 Nov 2009 15:54:39 +0100 Message-ID: <4AFAD02F.9080305@mocean-labs.com> References: <4AFACC6A.304@mocean-labs.com> <20091111144924.GG30489@zod.rchland.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Cc: spi-devel-general@lists.sourceforge.net, Andrew Morton , dbrownell@users.sourceforge.net, John Linn , linuxppc-dev@ozlabs.org To: Josh Boyer Return-path: In-Reply-To: <20091111144924.GG30489@zod.rchland.ibm.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: linuxppc-dev-bounces+glppd-linuxppc64-dev=m.gmane.org@lists.ozlabs.org Errors-To: linuxppc-dev-bounces+glppd-linuxppc64-dev=m.gmane.org@lists.ozlabs.org List-Id: linux-spi.vger.kernel.org Josh Boyer wrote: > On Wed, Nov 11, 2009 at 03:38:34PM +0100, Richard R=F6jfors wrote: >> This patch changes the out_(be)(8|16|32) and in_(be)(8|16|32) calls to i= owrite(8|16|32) >> and ioread(8|16|32). This to be able to build on platforms not supportin= g the in/out calls >> for instance x86. >> >> Support is also added for little endian writes. In some systems the regi= sters should be >> accessed little endian rather than big endian. > = > I wonder if you should make the endianness a config option. Right now you > have a conditional check for every read and write. Does that impact > performance at all? It won't affect the performance noticeable. In our case we need the possibility to run both endians using the same xili= nx_spi module. --Richard