Re: [PATCH v2 0/3] rtc: ds1685: power-off fixes

[Joshua Kinard <linux@kumba.dev>]

On 10/13/2026 08:02, Imre Kaloz wrote:
> Three fixes for ds1685_rtc_poweroff(), on SGI O2 and Octane machines.
> 
> Patch 1 arms the wake-up alarm only if an alarm is set, so that stale
> alarm registers no longer power the machine back on.
> 
> Patch 2 replaces the sleep before the PAB write with a busy wait, as
> power-off handlers must not sleep.
> 
> Patch 3 repeats the PAB write, clearing the kickstart and wake-up
> flags with it, until power is removed.
> 
> Patch 1 was tested on an O2, and the whole series on an Octane.
> 
> Changes in v2:
> - New patch 2: use mdelay() instead of msleep() before the PAB write.
> - Patch 3 (v1 patch 2): use mdelay() instead of msleep() in the
>    repeat loop.
> 
> v1: https://lore.kernel.org/all/20261013094930.574-1-kaloz@kernel.org/
> 
> Imre Kaloz (3):
>    rtc: ds1685: arm the wake-up alarm at power-off only if an alarm is
>      set
>    rtc: ds1685: do not sleep in the power-off handler
>    rtc: ds1685: keep setting PAB until power is removed
> 
>   drivers/rtc/rtc-ds1685.c | 40 ++++++++++++++++++++++++++--------------
>   1 file changed, 26 insertions(+), 14 deletions(-)
> 
> 
> base-commit: 24a19b953e2a3d0ea5a8c4951a0ab8f32720c907

Whoops, I meant to ACK this thread, not the older one!

Acked-by: Joshua Kinard <linux@kumba.dev>