diff options
author | Rafael J. Wysocki <rafael.j.wysocki@intel.com> | 2020-02-11 10:11:02 +0100 |
---|---|---|
committer | Rafael J. Wysocki <rafael.j.wysocki@intel.com> | 2020-02-11 10:11:02 +0100 |
commit | e3728b50cd9be7d4b1469447cdf1feb93e3b7adb (patch) | |
tree | 446902616310615718c4615e9dbe1a1ea2e5a914 /drivers/acpi | |
parent | f0ac20c3f6137910c8a927953e8a92f5b3716166 (diff) | |
download | linux-e3728b50cd9be7d4b1469447cdf1feb93e3b7adb.tar.gz linux-e3728b50cd9be7d4b1469447cdf1feb93e3b7adb.tar.bz2 linux-e3728b50cd9be7d4b1469447cdf1feb93e3b7adb.zip |
ACPI: PM: s2idle: Avoid possible race related to the EC GPE
It is theoretically possible for the ACPI EC GPE to be set after the
s2idle_ops->wake() called from s2idle_loop() has returned and before
the subsequent pm_wakeup_pending() check is carried out. If that
happens, the resulting wakeup event will cause the system to resume
even though it may be a spurious one.
To avoid that race, first make the ->wake() callback in struct
platform_s2idle_ops return a bool value indicating whether or not
to let the system resume and rearrange s2idle_loop() to use that
value instad of the direct pm_wakeup_pending() call if ->wake() is
present.
Next, rework acpi_s2idle_wake() to process EC events and check
pm_wakeup_pending() before re-arming the SCI for system wakeup
to prevent it from triggering prematurely and add comments to
that function to explain the rationale for the new code flow.
Fixes: 56b991849009 ("PM: sleep: Simplify suspend-to-idle control flow")
Cc: 5.4+ <stable@vger.kernel.org> # 5.4+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Diffstat (limited to 'drivers/acpi')
-rw-r--r-- | drivers/acpi/sleep.c | 44 |
1 files changed, 31 insertions, 13 deletions
diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c index 439880629839..2c695b196cd2 100644 --- a/drivers/acpi/sleep.c +++ b/drivers/acpi/sleep.c @@ -990,21 +990,28 @@ static void acpi_s2idle_sync(void) acpi_os_wait_events_complete(); /* synchronize Notify handling */ } -static void acpi_s2idle_wake(void) +static bool acpi_s2idle_wake(void) { - /* - * If IRQD_WAKEUP_ARMED is set for the SCI at this point, the SCI has - * not triggered while suspended, so bail out. - */ - if (!acpi_sci_irq_valid() || - irqd_is_wakeup_armed(irq_get_irq_data(acpi_sci_irq))) - return; + if (!acpi_sci_irq_valid()) + return pm_wakeup_pending(); + + while (pm_wakeup_pending()) { + /* + * If IRQD_WAKEUP_ARMED is set for the SCI at this point, the + * SCI has not triggered while suspended, so bail out (the + * 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))) + return true; + + /* + * If there are no EC events to process, the wakeup is regarded + * as a genuine one. + */ + if (!acpi_ec_dispatch_gpe()) + return true; - /* - * If there are EC events to process, the wakeup may be a spurious one - * coming from the EC. - */ - if (acpi_ec_dispatch_gpe()) { /* * Cancel the wakeup and process all pending events in case * there are any wakeup ones in there. @@ -1017,8 +1024,19 @@ static void acpi_s2idle_wake(void) acpi_s2idle_sync(); + /* + * The SCI is in the "suspended" state now and it cannot produce + * new wakeup events till the rearming below, so if any of them + * are pending here, they must be resulting from the processing + * of EC events above or coming from somewhere else. + */ + if (pm_wakeup_pending()) + return true; + rearm_wake_irq(acpi_sci_irq); } + + return false; } static void acpi_s2idle_restore_early(void) |