On 04/01/2016 05:33 PM, Mathieu Poirier wrote: > On 1 April 2016 at 08:22, Neil Armstrong <narmstr...@baylibre.com> wrote: Hi Matthieu, Thanks for the review.
>> + /* OX810SE Uses a special 24bit width */ >> + if (of_device_is_compatible(np, "oxsemi,ox810se-rps-timer")) >> + width = 24; > > Wouldn't it be better to add a new optional property for this? If the > property is not specified then we default to a width of 32. Otherwise > the new width is configured. That way we don't have to add a new > compatible string every time (however often) a special width is > encountered. As explained by Rob Herring, this HW is compatible with SP804 but is /not/ an IP from ARM and has been customized to support 24bit. Having a DT property for this width is not wanted actually since it makes no sense for the original SP804 IP, so it should be handled per vendor. > > Otherwise I think the code facelift looks good, > Mathieu Is there any other blocking points ? The point to have a change on sp804 is to economize a new very similar driver for oxnas. Neil
