A number of users have reported their system will have a failure reading the RTC around s2idle entry or exit. This failure manifests as UIP clear taking longer than 10ms. Affected users have reported that after this happens the clock jumps forward to 2077, which is presumably from epoch + century bit. Users who used a debugging patch provided by Mateusz Jończyk demonstrated that this has taken upwards of 480ms in some cases. This series adjusts the UIP timeout to be configurable by the caller and changes some callers which aren't called in an interrupt context to allow longer timeouts. Link: https://bugzilla.kernel.org/show_bug.cgi?id=217626 Link: https://community.frame.work/t/responded-fw13-amd-fedora-39-system-clock-advances-50-years-during-overnight-suspend Mario Limonciello (4): rtc: mc146818-lib: Adjust failure return code for mc146818_get_time() rtc: Adjust failure return code for cmos_set_alarm() rtc: Add support for configuring the UIP timeout for RTC reads rtc: Extend timeout for waiting for UIP to clear to 1s arch/alpha/kernel/rtc.c | 2 +- arch/x86/kernel/hpet.c | 2 +- arch/x86/kernel/rtc.c | 2 +- drivers/base/power/trace.c | 2 +- drivers/rtc/rtc-cmos.c | 8 ++++---- drivers/rtc/rtc-mc146818-lib.c | 35 ++++++++++++++++++++++++---------- include/linux/mc146818rtc.h | 3 ++- 7 files changed, 35 insertions(+), 19 deletions(-) -- 2.34.1
