From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Brown Subject: Re: [PATCHv4 4/6] input/cma3000_d0x: Add CMA3000 spi support Date: Tue, 25 Oct 2011 10:36:34 +0200 Message-ID: <20111025083633.GB31508@opensource.wolfsonmicro.com> References: <1319487676-27181-1-git-send-email-ricardo.ribalda@gmail.com> <1319487676-27181-5-git-send-email-ricardo.ribalda@gmail.com> <20111024213635.GB4394@ponder.secretlab.ca> <4EA6741B.6090508@cam.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <4EA6741B.6090508@cam.ac.uk> Sender: linux-kernel-owner@vger.kernel.org To: Jonathan Cameron Cc: Ricardo Ribalda Delgado , Grant Likely , dmitry.torokhov@gmail.com, sameo@linux.intel.com, peter.ujfalusi@ti.com, aghayal@codeaurora.org, david@hardeman.nu, Shubhrajyoti@ti.com, saaguirre@ti.com, hemanthv@ti.com, linux-input@vger.kernel.org, LKML List-Id: linux-input@vger.kernel.org On Tue, Oct 25, 2011 at 09:32:27AM +0100, Jonathan Cameron wrote: > On 10/25/11 08:24, Ricardo Ribalda Delgado wrote: > > It is not on linux-next, and also it cannot be used directly, in order > > to read you have to do set reg= (reg << 2) | 2; And AFAIK it is not > > supported by the regmap > My gut feeling is it would easy enough to add if it can't currently > be done. Mark? That should be trivial to add - the | 2 bit is already supported in -next and will be sent to Linus this merge window, adding the shift is just a simple matter of programming. > Gains here are leveraging the bus abstractions from there. There aren't > enough registers to gain from caching etc. Depends on the workload more than the number of registers - if you're able to eliminate enough reads then it can give a bit of a performance boost. But it's completely optional either way.