On Wed, Nov 19, 2014 at 11:45:14PM -0800, Dmitry Torokhov wrote: > On Wed, Nov 19, 2014 at 09:15:41PM -0800, Doug Anderson wrote: > > I will defer to your wisdom here. I agree that these are the two > > primary solutions and I've picked one, but I have no idea which will > > be more of a PITA in the long run. > > > > Note: I'm not sure that anyone expects EPROBE_DEFER to be returned > > from a clk_enable() (do they?). It almost seems like the right answer > > is to fail to allow anyone to clk_get() any clock that doesn't have a > > path to root. > > EPROBE_DEFER only makes sense in driver's probe paths and so I would be > very against adding it to clk_enable() which is called from many places > in the kernel. If we decide to go with EPROBE_DEFER then returning it > from clk_get() seems like a much better choice since it is normally > called during probing. Absolutely correct. EINVAL would be better for clk_prepare() since it isn't something that can be recovered from by just retrying a bit later. You're absolutely correct that EPROBE_DEFER has no business being returned in any path other than a driver's probe function; it is not a user visible error code, it is a special internal Linux error code which only the driver model understands to mean "add this device to the deferred probe list and try again a while later." Userspace, especially, should never see this error code. -- FTTC broadband for 0.8mile line: currently at 9.5Mbps down 400kbps up according to speedtest.net.
