Hello Tom, Jérôme,
Sorry for the late reply, but I just starting to look through my backlog, after my vacation ...
Tom Rini wrote:
On Wed, Dec 21, 2011 at 10:36 PM, Jérôme Carretero cJ-uboot@zougloub.eu wrote:
Hi Tom,
I'm requesting comments on the following (untested) patch. It adds support for multiple i2c buses on davinci_i2c, without altering one line of code. What I don't like is that I'm doing a bit of macro black magic to transform constants into variables. But I don't know whether refactoring the code would be accepted, so this is a first shot.
So, this follows the example of other i2c drivers, but as Wolfgang likes to point out, following another bad example isn't a valid reason to do something. So, is this the right way or a bad example to follow, Wolfgang or Heiko? Thanks!
Yes, that is the actual (not really nice) state to implement the "multibus" feature for i2c. The better way would be to go this way:
http://git.denx.de/?p=u-boot/u-boot-i2c.git;a=shortlog;h=refs/heads/multibus...
but this branch is not up to date, some test on at last powerpc and arm architectures should be done, and there are a lof of checkpatch errors in this branch (any help is welcome ;-) ... so there is some work to do, before this can go to mainline ...
bye, Heiko