platform/x86/intel: hibernation aborts due to spurious power button event
Subsystems (2)
cover platform/x86/intel: hibernation aborts due to spurious power button event
Last year I submitted a patch that partially addressed the issues described in:
https://bugzilla.kernel.org/show_bug.cgi?id=218634
I recently got a Dell Precision 3680 workstation, and came across the problem that's still affecting some users in that thread: when hibernating (even from software), a 0xCE (power button down) event is received, and hibernation is aborted.
I've attached a patch that I'm currently testing. I think it makes sense based on the pm.h description of freeze:
Quoted 2 lines
Analogous to @suspend(), but it should not enable the device to signal wakeup events or change its power state
However, I'm not sure about the implementation. Perhaps it would be better to set wakeup_mode in suspend instead of prepare? As it is, wakeup_mode will be set between prepare and freeze, which feels wrong.
I also did some investation into the source of the spurious events, but I'm a little out of my depth.
A normal wakeup from suspend with the power button looks like this:
[ 1549.112851] intel-hid INTC1070:00: pm_prepare wakeup_mode 1 [ 1549.117033] serial 00:03: disabled [ 1549.173790] e1000e: EEE TX LPI TIMER: 00000011 [ 1549.363439] ACPI: EC: interrupt blocked [ 1559.055878] evmisc-0132 ev_queue_notify_reques: Dispatching Notify on [HIDD] (Device) Value 0xCE (Hardware-Specific) Node 00000000a22ff378 [ 1559.056071] intel-hid INTC1070:00: event 0xce wakeup_mode 1 [ 1559.059997] ACPI: EC: interrupt unblocked [ 1559.736517] intel-hid INTC1070:00: pm_complete wakeup_mode 0 [ 1559.747551] i915 0000:00:02.0: [drm] GT0: GuC firmware i915/tgl_guc_70.bin version 70.44.1 [ 1559.747555] i915 0000:00:02.0: [drm] GT0: HuC firmware i915/tgl_huc.bin version 7.9.3 [ 1559.749329] spd5118 0-0051: PM: dpm_run_callback(): spd5118_resume [spd5118] returns -6 [ 1559.749334] spd5118 0-0051: PM: failed to resume async: error -6 [ 1559.749627] spd5118 0-0053: PM: dpm_run_callback(): spd5118_resume [spd5118] returns -6 [ 1559.749631] spd5118 0-0053: PM: failed to resume async: error -6 [ 1559.749996] i915 0000:00:02.0: [drm] GT0: HuC: authenticated for all workloads [ 1559.750719] i915 0000:00:02.0: [drm] GT0: GUC: submission enabled [ 1559.750720] i915 0000:00:02.0: [drm] GT0: GUC: SLPC enabled [ 1559.751300] i915 0000:00:02.0: [drm] GT0: GUC: RC enabled [ 1559.751343] evmisc-0132 ev_queue_notify_reques: Dispatching Notify on [HIDD] (Device) Value 0xCF (Hardware-Specific) Node 00000000a22ff378 [ 1559.751396] intel-hid INTC1070:00: event 0xcf wakeup_mode 0
The spurious event in hibernation looks like:
[ 2510.569198] ACPI: PM: Waking up from system sleep state S4 [ 2510.570714] evmisc-0132 ev_queue_notify_reques: Dispatching Notify on [RP02] (Device) Value 0x00 (Bus Check) Node 0000000074f801ab [ 2510.570775] evmisc-0132 ev_queue_notify_reques: Dispatching Notify on [RP05] (Device) Value 0x00 (Bus Check) Node 00000000bb982283 [ 2510.571061] evmisc-0132 ev_queue_notify_reques: Dispatching Notify on [RP21] (Device) Value 0x00 (Bus Check) Node 000000009ba3a342 [ 2510.575365] evmisc-0132 ev_queue_notify_reques: Dispatching Notify on [HIDD] (Device) Value 0xCE (Hardware-Specific) Node 00000000a22ff378 [ 2510.575369] intel-hid INTC1070:00: event 0xce wakeup_mode 1 [ 2510.575715] ACPI: EC: interrupt unblocked [ 2510.602774] ACPI: EC: event unblocked [ 2510.602887] intel-hid INTC1070:00: pm_complete wakeup_mode 0
There's some code at the end of acpi_pm_finish which sends a power button event, but it doesn't seem to be hitting that.
David McFarland (1):
platform/x86/intel: disable wakeup_mode during hibernation
drivers/platform/x86/intel/hid.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)
Signature · 1 line
2.50.1
PATCH 1/1 platform/x86/intel: disable wakeup_mode during hibernation
--- drivers/platform/x86/intel/hid.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)
drivers/platform/x86/intel/hid.c +9 -1
@@ -406,6 +406,14 @@ static int intel_hid_pl_suspend_handler(struct device *device)return 0 ; } +static int intel_hid_pl_freeze_handler (struct device *device) +{ +struct intel_hid_priv *priv =dev_get_drvdata (device); + + priv->wakeup_mode = false; +return intel_hid_pl_suspend_handler (device); +} +static int intel_hid_pl_resume_handler (struct device *device) {intel_hid_pm_complete (device);
@@ -420,7 +428,7 @@ static int intel_hid_pl_resume_handler(struct device *device)static const struct dev_pm_ops intel_hid_pl_pm_ops = { .prepare = intel_hid_pm_prepare, .complete = intel_hid_pm_complete, - .freeze = intel_hid_pl_suspend_handler, + .freeze = intel_hid_pl_freeze_handler, .thaw = intel_hid_pl_resume_handler, .restore = intel_hid_pl_resume_handler, .suspend = intel_hid_pl_suspend_handler,
Signature · 1 line
2.50.1