From mboxrd@z Thu Jan 1 00:00:00 1970 From: Adrian McMenamin Date: Sat, 20 Dec 2008 20:47:00 +0000 Subject: Re: [PATCH] sh: maple: add support for Maple controller as a Message-Id: <1229806020.6502.35.camel@localhost.localdomain> List-Id: References: <1229728542.1441.6.camel@localhost.localdomain> <20081220151511.GA6528@console-pimps.org> <1229789248.6502.4.camel@localhost.localdomain> <200812201219.12511.dmitry.torokhov@gmail.com> In-Reply-To: <200812201219.12511.dmitry.torokhov@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Dmitry Torokhov Cc: Matt Fleming , LKML , linux-sh , Andrew Morton , Paul Mundt On Sat, 2008-12-20 at 12:19 -0800, Dmitry Torokhov wrote: > On Saturday 20 December 2008 08:07:27 Adrian McMenamin wrote: > > On Sat, 2008-12-20 at 15:16 +0000, Matt Fleming wrote: > > > On Fri, Dec 19, 2008 at 11:15:42PM +0000, Adrian McMenamin wrote: > > > > Add support for the SEGA Dreamcast Maple controller as a joystick > > > > > > [snip] > > > > > > > + struct dc_pad *pad = maple_get_drvdata(mapledev); > > > > + struct input_dev *dev = pad->dev; > > > > + unsigned char *res = mq->recvbuf; > > > > + > > > > + buttons = ~cpu_to_le16(*(unsigned short *)(res + 8)); > > > > + > > > > > > I may be wrong but shouldn't this use the I/O accessor functions? > > > > It's not iomemory, so I don't think so. The maple bus writes to a buffer > > in the physical RAM. > > > > Hmm, but why we are going _from_ cpu order to le16??? I'd expect we > wanted to do conversion the other way around. I pulled it out of 'next' > for now. > The data is transmitted here by the bus in 32 bit little endian blocks but makes sense in le16 bit lumps. Off is 1 and on is 0 - hence the ~ The first 16 bits are the button data the next 16 bits are the (start of) the axis data. Both cpu and bus data (in this case - it can vary) are little endian.