Hi, Thanks for the report. On 06/07/2017 at 15:16:11 -0400, elseifthen@xxxxxxx wrote: > There appears to be a hardware specific rtc regression in v4.4.76. > > Introduced in v4.4.52 by: > > commit f0414c1f8bb7a4e69064296f460773170c5435ac > Author: Colin Ian King <colin.king@xxxxxxxxxxxxx> > Date: Mon May 16 17:22:54 2016 +0100 > > rtc: interface: ignore expired timers when enqueuing new timers > > commit 2b2f5ff00f63847d95adad6289bd8b05f5983dd5 upstream. > > 8< --- > > The problem does not appear to be in v4.9.x kernels. > > Maybe v4.4.76 is missing one of these? > http://marc.info/?l=linux-rtc&m=149489307123605&w=2 > Which also includes the commit in question. > > See this thread for more details: > https://www.linuxquestions.org/questions/slackware-14/slow-hwclock-systohc-on-reboot-shutdown-4175609040/ > > Note: I am not the one having this issue. I'm only reporting it, so any > questions would need to be asked in the above thread. Or perhaps the > person with the problem will follow this email thread. I will post a > link to it for them. > I can see a small issue with f0414c1f8bb7a4e69064296f460773170c5435ac but I'm not sure why it would be hit on v4.4 and not v4.9. Maybe one has CONFIG_RTC_INTF_DEV_UIE_EMUL and not the other one. Could you try with the following on top of v4.4.57 (or any other v4.4.x): --- a/drivers/rtc/interface.c +++ b/drivers/rtc/interface.c @@ -764,7 +764,7 @@ static int rtc_timer_enqueue(struct rtc_device *rtc, struct rtc_timer *timer) } timerqueue_add(&rtc->timerqueue, &timer->node); - if (!next) { + if (!next || ktime_before(timer->node.expires, next->expires)) { struct rtc_wkalrm alarm; int err; alarm.time = rtc_ktime_to_tm(timer->node.expires); -- Alexandre Belloni, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com
