On Thu, Aug 17, 2023 at 09:51:17PM +0200, Alexandre Belloni wrote: > On 16/08/2023 12:12:14-0700, Guenter Roeck wrote: > > On Wed, Aug 16, 2023 at 06:14:35PM +0200, Alexandre Belloni wrote: > > > On 16/08/2023 08:50:12-0700, Guenter Roeck wrote: > > > > > I'm fine with the series, however, this doesn't solve the issue for RTCs > > > > > that have an absolute limit on the alarm (as opposed to an offset to the > > > > > current time/date). > > > > > > > > > > > > > I thought that is checked by rtc_valid_range() in rtc_set_alarm(). > > > > Am I missing something ? Of course that assumes that the absolute > > > > maximum alarm timeout matches range_max, but I didn't find any > > > > drivers where that would not be the case. > > > > > > > > > > There are RTCs where this is not the case. When this is far away in the > > > future enough, the usual solution is to clip range_max which works but > > > is not really great intellectually. > > > > > Do you have an example, by any chance ? > > > > I'm sorry, I've been looking and I couldn't find it anymore. Don't > bother with this for now. > I have been looking as well and did not find it either. I'll go ahead and send v2 with the suggested changes. Thanks, Guenter
