On Wed, 20 May 2026, Mario Limonciello wrote:

> It was reported that suspend-then-hibernate stopped working with modern
> systemd versions on AMD Cezanne systems. The reason for this breakage
> was because systemd switched to using alarmtimer instead of the wakealarm
> sysfs file.
> 
> On AMD Cezanne systems, amd_pmc_verify_czn_rtc() programs a secondary
> timer with the alarm time. This was introduced by
> commit 59348401ebed ("platform/x86: amd-pmc: Add special handling for
> timer based S0i3 wakeup"). However, this function uses rtc_read_alarm(),
> which only reads the aie_timer, not the next expiring timer from the
> timerqueue.
> 
> When both alarmtimer and wakealarm are active, the first expiring timer
> might be the alarmtimer, but amd_pmc_verify_czn_rtc() would only see
> the aie_timer, potentially missing the earlier alarm.
> 
> Switch to rtc_read_next_alarm() to read whichever timer will fire next.
> Also handle -ENOENT (no alarm pending) explicitly as a non-error case.
> 
> Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/3591
> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
> ---
>  drivers/platform/x86/amd/pmc/pmc.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/platform/x86/amd/pmc/pmc.c b/drivers/platform/x86/amd/pmc/pmc.c
> index 50f5784f2aa2e..8cd2db0ccaacd 100644
> --- a/drivers/platform/x86/amd/pmc/pmc.c
> +++ b/drivers/platform/x86/amd/pmc/pmc.c
> @@ -595,9 +595,12 @@ static int amd_pmc_verify_czn_rtc(struct amd_pmc_dev *pdev, u32 *arg)
>  	rtc_device = rtc_class_open("rtc0");
>  	if (!rtc_device)
>  		return 0;
> -	rc = rtc_read_alarm(rtc_device, &alarm);
> -	if (rc)
> -		return rc;
> +	rc = rtc_read_next_alarm(rtc_device, &alarm);
> +	if (rc) {
> +		if (rc == -ENOENT)
> +			dev_dbg(pdev->dev, "no alarm pending\n");
> +		return rc == -ENOENT ? 0 : rc;
> +	}
>  	if (!alarm.enabled) {
>  		dev_dbg(pdev->dev, "alarm not enabled\n");
>  		return 0;

Acked-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>

-- 
 i.