On Wed, Oct 29, 2014 at 08:36:41AM -0700, Guenter Roeck wrote: > On Wed, Oct 29, 2014 at 02:35:26PM +0100, Johan Hovold wrote: > > On Wed, Oct 29, 2014 at 06:20:40AM -0700, Guenter Roeck wrote: > > > On 10/29/2014 05:34 AM, Johan Hovold wrote: > > > > And what about any power-off latencies? Should this always be dealt with > > > > in the power-off handler? > > > > > > > > Again, if it's predictable and high, as in the OMAP RTC case, it should > > > > go in the handler. But what if it's just normal bus latencies > > > > (peripheral busses, i2c, or whatever people may come up with)? > > > > > > > > Should there always be a short delay before calling the next handler? > > > > > > That delay would depend on the individual power-off handler, so I think > > > the current implementation works just fine (where power-off handlers > > > implement the delay). > > > > Some don't, and could possibly unknowingly have been relying on the fact > > that they could return to user space and be powered off at some later > > time. With systemd that would have caused a panic. > > Agreed, but there are two cases to consider: What should be the delay > before the next power-off handler is called, and what should the system > do if all power-off handlers fail (or if there are none). The current > behavior isn't exactly well defined. Ok, with systemd that results in > a crash, but I am not really sure if one can or should blame systemd > for that. The discussion about systemd and its philosophy should not > cloud the fact that power-off behavior isn't exactly well defined. Sounds like we pretty much agree. See my response to your last mail. > > Also consider generic power-off handlers such as gpio-poweroff. It > > currently hard-codes a three-second delay but the actual delay would > > really be board specific. > > > A configurable delay would address that. The actually required delay > could be provided in platform data or as devicetree property. Yep, see mail mentioned above. Johan -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
