On Sun, 7 Nov 2010, Philip Rakity wrote: > > On Nov 6, 2010, at 6:44 PM, Nicolas Pitre wrote: > > > On Sat, 6 Nov 2010, Philip Rakity wrote: > > > >> > >> We need a way to specify > >> > >> a) Hardware can support clock gating setting the clock to 0 in the > >> call to set_ios (with hardware clock gating enabled) becomes a no-op > >> so some care is needed which is why I suggested another field in the > >> ios structure. > > > > Why should the core code care? > > > > What difference is there with the core telling the controller driver to > > set the clock to a frequency of 0 ? > > When h/w clock gating is enabled then sd clocks are disabled/enabled > by the sd controller; there is nothing for the core/ layer to do. OK, then that could be an optimization. > I am concerned that having the core/ layer ask for clocks off and on > is > > a) not needed when h/w clock gating is enabled Sure, but right now it shouldn't be enabled just yet. Making sure that the software version works well should be done first. > b) may invoke bad behavior in the controller That would be a really bad controller. > c) is not optimal since the h/w is already handling this. -- Agreed. But again let's make sure that the software version works first. Then adding support for hardware clock gating would be really simple. Nicolas -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
