On Wed, Aug 08, 2018 at 09:59:28AM +0200, Wolfram Sang wrote:
> When doing a REP_START after a read message, the driver used to trigger
> a STOP first which would then be overwritten by REP_START. This was the
> only stable method found when doing the last refactoring. However, this
> was not in accordance with the documentation.
> 
> After research from our BSP team and myself, we now can implement a
> version which works and is according to the documentation. The new
> approach ensures the ICMCR register is only changed when really needed.
> 
> Tested on a R-Car Gen2 (H2) and Gen3 with DMA (M3N).
> 
> Signed-off-by: Hiromitsu Yamasaki <[email protected]>
> Signed-off-by: Wolfram Sang <[email protected]>

Applied to for-next, thanks!

Attachment: signature.asc
Description: PGP signature

Reply via email to