On 13/10/2018 14.15, Jonathan Cameron wrote: > On Tue, 9 Oct 2018 12:57:07 +0100 > Chris Coffey <cmc@xxxxxxxxxxxxx> wrote: > >> Hi Jonathan, thank you for the review. I added a few comments below >> based on your feedback. >> >> - Chris >> >> On Mon, Oct 08, 2018 at 09:08:29PM +0100, Jonathan Cameron wrote: >>> That's interesting. I'm not sure we have any consistency of interface >>> around whether a write to the value when powered down results in the device >>> powering up or not. It certainly feels like we should be consistent on this >>> and document it. >>> >>> I checked the first random driver I found the ad5360 and it appears to >>> have the powerdown on the front end in some sense in that there is a >>> specific bit to clear in order to power up again and it is not done >>> by changing the current value. >>> >>> To my mind that is the more logical option, but I'd like others opinions >>> on this. >>> >> I see what you mean. My intent was to mirror the behavior of userspace >> programs and APIs that I've written and seen for these chips. They take >> advantage of the fact a single command to the chip sets the voltage >> level and power state. That way there is no need to issue a separate >> power-up command, though it's available in sysfs if a user wants it >> (e.g., to toggle the channel's power state without specifying a new >> voltage level). >> >> Like you, I'd be curious to hear what others have to say. My DAC >> experience is almost exclusively with Microchip devices, so I may be >> suffering from a bit of tunnel vision here. > > Agreed. I've added a few additional CCs. If we don't get any > replies we may want to raise a specific RFC email on this to catch > people's attention. > > The cc list is fairly random, so if people wouldn't mind drawing > the attention of others to this question that would be very helpful! > > Jonathan > Hi The DAC5571 is like this one able to leave powerdown mode with a single write command. But I'm blocking from using it in the driver so we are consistent across all the DAC devices. (Hopefully) For me it makes sense to use the powerdown API to control all powerdown settings... /Sean
