Re: [PATCH 04/05] input: RMI4 F01 device control

On 01/31/2013 01:14 PM, Christopher Heiny wrote:

On 01/31/2013 12:08 AM, Dmitry Torokhov wrote:

quoted

Hi Chris,

On Fri, Jan 18, 2013 at 05:12:44PM -0800, Christopher Heiny wrote:

quoted

In addition to the changes described in 0/0 of this patchset, this patch
includes device serialization updated to conform to the latest RMI4
specification.

I was looking at the various aspects of the RMI4 patchset, trying to
fix the issues that I see, but there is one big issue that I simply do
not have time to tackle - the driver is completely broken on big endian
architectures due to reliance on bitfileds when exchanging the data with
the device.

Grumble.  I brought this up during the original development, but was
assured that it would work fine per the C language spec.  I'll have to
revisit that discussion and report back.

OK, after some trivial research it turns out you're right, and I was right originally, but let myself be blindsided by bafflegab.  Doh!

As you mentioned, fixing the endian-ness problem is a pretty wide ranging change to the way things are done in the driver.  Fortunately it's not a structural change to the architecture of the driver.

But because it affects pretty much the whole code (which is actually a LOT more code than we're currently submitting - but that unsubmitted code will have to be updated concurrently with the submitted code to prevent a trainwreck), I'd like to isolate that fix to a single event.

So I propose we get all our other issues worked out, since a bunch of them are already in progress.  Once that stuff is stable, my team can do one big change to fix the endian bug.  If the code is stable, we can do the update on our end pretty quickly.

How does that sound?

					Chris