https://bugzilla.novell.com/show_bug.cgi?id=884899 https://bugzilla.novell.com/show_bug.cgi?id=884899#c10 --- Comment #10 from Borislav Petkov <bpetkov@suse.com> 2014-07-07 14:31:05 UTC --- Hmm, so your second dmesg looks interesting: [ 583.298388] hpet_rtc_timer_reinit: 84 callbacks suppressed [ 583.298391] hpet1: lost 1099 rtc interrupts and since rtc is using hpet interrupts: [ 1.099227] rtc_cmos 00:01: alarms up to one month, y3k, 114 bytes nvram, hpet irqs ^^^^^^^^^ I can imagine us losing some of the hpet interrupts during suspend/resume could cause us to miss out on the wakeup alarm. And AFAIR, on those boxes/chipsets, HPET is emulated in the BIOS so it could be that the HPET might have problems if the machine is suspended. Now, I've checked whether upstream has gotten any new fixes to hpet.c or rtc-cmos.c since 13.1 but I don't see anything relevant. You said above:
Workaround: install kernel-desktop-3.7.10-1.32.1 from opensuse-12.3
which would mean that some commit after 3.7 introduced the regression. So the only thing I could think of right now is to bisect it. Basically grab an upstream kernel git tree and first build and test in a coarse-grained manner kernels 3.7, 3.8, ... 3.15. At the first major release number X which shows the regression, you do a normal bisect between kernels 3.X-1 and 3.X as we're assuming 3.X-1 was still ok. I know, it is time consuming so it is up to you. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.