Merge tag 'pm-5.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull power management updates from Rafael Wysocki:
 "These rework the system-wide PM driver flags, make runtime switching
  of cpuidle governors easier, improve the user space hibernation
  interface code, add intel-speed-select interface documentation, add
  more debug messages to the ACPI code handling suspend to idle, update
  the cpufreq core and drivers, fix a minor issue in the cpuidle core
  and update two cpuidle drivers, improve the PM-runtime framework,
  update the Intel RAPL power capping driver, update devfreq core and
  drivers, and clean up the cpupower utility.

  Specifics:

   - Rework the system-wide PM driver flags to make them easier to
     understand and use and update their documentation (Rafael Wysocki,
     Alan Stern).

   - Allow cpuidle governors to be switched at run time regardless of
     the kernel configuration and update the related documentation
     accordingly (Hanjun Guo).

   - Improve the resume device handling in the user space hibernarion
     interface code (Domenico Andreoli).

   - Document the intel-speed-select sysfs interface (Srinivas
     Pandruvada).

   - Make the ACPI code handing suspend to idle print more debug
     messages to help diagnose issues with it (Rafael Wysocki).

   - Fix a helper routine in the cpufreq core and correct a typo in the
     struct cpufreq_driver kerneldoc comment (Rafael Wysocki, Wang
     Wenhu).

   - Update cpufreq drivers:

      - Make the intel_pstate driver start in the passive mode by
        default on systems without HWP (Rafael Wysocki).

      - Add i.MX7ULP support to the imx-cpufreq-dt driver and add
        i.MX7ULP to the cpufreq-dt-platdev blacklist (Peng Fan).

      - Convert the qoriq cpufreq driver to a platform one, make the
        platform code create a suitable device object for it and add
        platform dependencies to it (Mian Yousaf Kaukab, Geert
        Uytterhoeven).

      - Fix wrong compatible binding in the qcom driver (Ansuel Smith).

      - Build the omap driver by default for ARCH_OMAP2PLUS (Anders
        Roxell).

      - Add r8a7742 SoC support to the dt cpufreq driver (Lad
        Prabhakar).

   - Update cpuidle core and drivers:

      - Fix three reference count leaks in error code paths in the
        cpuidle core (Qiushi Wu).

      - Convert Qualcomm SPM to a generic cpuidle driver (Stephan
        Gerhold).

      - Fix up the execution order when entering a domain idle state in
        the PSCI driver (Ulf Hansson).

   - Fix a reference counting issue related to clock management and
     clean up two oddities in the PM-runtime framework (Rafael Wysocki,
     Andy Shevchenko).

   - Add ElkhartLake support to the Intel RAPL power capping driver and
     remove an unused local MSR definition from it (Jacob Pan, Sumeet
     Pawnikar).

   - Update devfreq core and drivers:

      - Replace strncpy() with strscpy() in the devfreq core and use
        lockdep asserts instead of manual checks for a locked mutex in
        it (Dmitry Osipenko, Krzysztof Kozlowski).

      - Add a generic imx bus scaling driver and make it register an
        interconnect device (Leonard Crestez, Gustavo A. R. Silva).

      - Make the cpufreq notifier in the tegra30 driver take boosting
        into account and delete an unuseful error message from that
        driver (Dmitry Osipenko, Markus Elfring).

   - Remove unneeded semicolon from the cpupower code (Zou Wei)"

* tag 'pm-5.8-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (51 commits)
  cpuidle: Fix three reference count leaks
  PM: runtime: Replace pm_runtime_callbacks_present()
  PM / devfreq: Use lockdep asserts instead of manual checks for locked mutex
  PM / devfreq: imx-bus: Fix inconsistent IS_ERR and PTR_ERR
  PM / devfreq: Replace strncpy with strscpy
  PM / devfreq: imx: Register interconnect device
  PM / devfreq: Add generic imx bus scaling driver
  PM / devfreq: tegra30: Delete an error message in tegra_devfreq_probe()
  PM / devfreq: tegra30: Make CPUFreq notifier to take into account boosting
  PM: hibernate: Restrict writes to the resume device
  PM: runtime: clk: Fix clk_pm_runtime_get() error path
  cpuidle: Convert Qualcomm SPM driver to a generic CPUidle driver
  ACPI: EC: PM: s2idle: Extend GPE dispatching debug message
  ACPI: PM: s2idle: Print type of wakeup debug messages
  powercap: RAPL: remove unused local MSR define
  PM: runtime: Make clear what we do when conditions are wrong in rpm_suspend()
  Documentation: admin-guide: pm: Document intel-speed-select
  PM: hibernate: Split off snapshot dev option
  PM: hibernate: Incorporate concurrency handling
  Documentation: ABI: make current_governer_ro as a candidate for removal
  ...

1 files changed, 15 insertions, 5 deletions

diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c
index fd9d4e8318e9..31c9d0c8ae11 100644
--- a/drivers/acpi/sleep.c
+++ b/drivers/acpi/sleep.c
@@ -992,23 +992,31 @@ static bool acpi_s2idle_wake(void)
 		 * wakeup is pending anyway and the SCI is not the source of
 		 * it).
 		 */
-		if (irqd_is_wakeup_armed(irq_get_irq_data(acpi_sci_irq)))
+		if (irqd_is_wakeup_armed(irq_get_irq_data(acpi_sci_irq))) {
+			pm_pr_dbg("Wakeup unrelated to ACPI SCI\n");
 			return true;
+		}
 
 		/*
 		 * If the status bit of any enabled fixed event is set, the
 		 * wakeup is regarded as valid.
 		 */
-		if (acpi_any_fixed_event_status_set())
+		if (acpi_any_fixed_event_status_set()) {
+			pm_pr_dbg("ACPI fixed event wakeup\n");
 			return true;
+		}
 
 		/* Check wakeups from drivers sharing the SCI. */
-		if (acpi_check_wakeup_handlers())
+		if (acpi_check_wakeup_handlers()) {
+			pm_pr_dbg("ACPI custom handler wakeup\n");
 			return true;
+		}
 
 		/* Check non-EC GPE wakeups and dispatch the EC GPE. */
-		if (acpi_ec_dispatch_gpe())
+		if (acpi_ec_dispatch_gpe()) {
+			pm_pr_dbg("ACPI non-EC GPE wakeup\n");
 			return true;
+		}
 
 		/*
 		 * Cancel the SCI wakeup and process all pending events in case
@@ -1027,8 +1035,10 @@ static bool acpi_s2idle_wake(void)
 		 * are pending here, they must be resulting from the processing
 		 * of EC events above or coming from somewhere else.
 		 */
-		if (pm_wakeup_pending())
+		if (pm_wakeup_pending()) {
+			pm_pr_dbg("Wakeup after ACPI Notify sync\n");
 			return true;
+		}
 
 		rearm_wake_irq(acpi_sci_irq);
 	}