On Wed, 2014-11-12 at 07:03PM +0100, Andreas Färber wrote: > Am 12.11.2014 um 16:57 schrieb Sören Brinkmann: > > On Wed, 2014-11-12 at 02:51PM +0100, Andreas Färber wrote: > >> The specification requires xlnx,data-width, but example and driver use > >> xlnx,datawidth. Change the specification to match the implementation. > > > > Isn't this the wrong way around? The bindings are considered API, so > > shouldn't the driver be fixed to match the spec? > > In theory, patch review should've never let the two differ... ;) > > It's not my driver, so I fixed the perceived inconsistency the least > invasive way; Michal and Srikanth seemed to concur at the time. > > https://patchwork.kernel.org/patch/4620261/ > > > Are there already dts files out there using either of these options? > > In upstream, no. microblaze and virtex440 use a > xlnx,include-datawidth-matching-0 property as precedence for the > spelling, whereas there is an fsl,data-width and an unused msix-data-width. > > Downstream, yes: Beyond my own patch derived from the Parallella tree, > there's some in the ADI tree. None in the Xilinx tree on quick check. > > I haven't encountered any using the documented xlnx,data-width - but > this patch was authored pre 3.17, haven't ran a full Web search again. grepping through linux-next shows some usage of xlnx,datawidth, but only the single hit in Documentation for xlnx,data-width. Other than VDMA the other hit seems to be just some DT documentation which I can't find a driver for... Anyhow, this patch is probably the best way of fixing this. Reviewed-by: Soren Brinkmann <soren.brinkmann@xxxxxxxxxx> Thanks, Sören -- To unsubscribe from this list: send the line "unsubscribe dmaengine" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
