] ACPI: EC: s2idle: Avoid flushing EC work when EC GPE is inactive

From: Rafael J. Wysocki <[email protected]>

Flushing the EC work while suspended to idle when the EC GPE status is not set causes some EC wakeup events (notably power button and lid ones) to be missed after a series of spurious wakeups on the Dell XPS13 9360 in my office.

If that happens, the machine cannot be woken up from suspend-to-idle by a power button press or lid status change and it needs to be woken up in some other way (eg. by a key press).

Flushing the EC work only after successful dispatching the EC GPE, which means that its status has been set, avoids the issue, so change the code in question accordingly.

Signed-off-by Rafael J. Wysocki <[email protected]>

---

Hi Chris,

Please check if the key press wakeup still works on your system with this patch applied (on top of https://patchwork.kernel.org/patch/11538065/).

---
 drivers/acpi/ec.c    |    6 +++++-
 drivers/acpi/sleep.c |   15 ++++-----------
 2 files changed, 9 insertions(+), 12 deletions(-)

Index: linux-pm/drivers/acpi/ec.c
===================================================================

drivers/acpi/ec.c +5 -1

@@ -2020,9 +2020,13 @@ bool acpi_ec_dispatch_gpe(void)
         * to allow the caller to process events properly after that.
         */
        ret = acpi_dispatch_gpe(NULL, first_ec->gpe);
-       if (ret == ACPI_INTERRUPT_HANDLED)
+       if (ret == ACPI_INTERRUPT_HANDLED) {
                pm_pr_dbg("EC GPE dispatched\n");
 
+               /* Flush the event and query workqueues. */
+               acpi_ec_flush_work();
+       }
+
        return false;
 }
 #endif /* CONFIG_PM_SLEEP */

Index: linux-pm/drivers/acpi/sleep.c
===================================================================

drivers/acpi/sleep.c +4 -11

@@ -980,13 +980,6 @@ static int acpi_s2idle_prepare_late(void
        return 0;
 }
 
-static void acpi_s2idle_sync(void)
-{
-       /* The EC driver uses special workqueues that need to be flushed. */
-       acpi_ec_flush_work();
-       acpi_os_wait_events_complete(); /* synchronize Notify handling */
-}
-
 static bool acpi_s2idle_wake(void)
 {
        if (!acpi_sci_irq_valid())

@@ -1018,7 +1011,7 @@ static bool acpi_s2idle_wake(void)
                        return true;
 
                /*
-                * Cancel the wakeup and process all pending events in case
+                * Cancel the SCI wakeup and process all pending events in case
                 * there are any wakeup ones in there.
                 *
                 * Note that if any non-EC GPEs are active at this point, the

@@ -1026,8 +1019,7 @@ static bool acpi_s2idle_wake(void)
                 * should be missed by canceling the wakeup here.
                 */
                pm_system_cancel_wakeup();
-
-               acpi_s2idle_sync();
+               acpi_os_wait_events_complete();
 
                /*
                 * The SCI is in the "suspended" state now and it cannot produce

@@ -1060,7 +1052,8 @@ static void acpi_s2idle_restore(void)
         * of GPEs.
         */
        acpi_os_wait_events_complete(); /* synchronize GPE processing */
-       acpi_s2idle_sync();
+       acpi_ec_flush_work(); /* flush the EC driver's workqueues */
+       acpi_os_wait_events_complete(); /* synchronize Notify handling */
 
        s2idle_wakeup = false;
 

Quoted 16 lines · Rafael J. Wysocki

From: Rafael J. Wysocki <[email protected]>

Flushing the EC work while suspended to idle when the EC GPE status is not set causes some EC wakeup events (notably power button and lid ones) to be missed after a series of spurious wakeups on the Dell XPS13 9360 in my office.

If that happens, the machine cannot be woken up from suspend-to-idle by a power button press or lid status change and it needs to be woken up in some other way (eg. by a key press).

Flushing the EC work only after successful dispatching the EC GPE, which means that its status has been set, avoids the issue, so change the code in question accordingly.

Quoted 7 lines

Signed-off-by Rafael J. Wysocki <[email protected]>

---

Hi Chris,

Please check if the key press wakeup still works on your system with this patch applied (on top of https://patchwork.kernel.org/patch/11538065/).

Hi Chris,

Since I haven't heard back from you and the problem at hand is a regression on the machine where it happens, I'm going to push this patch for merging.

If it causes the key press wakeup issue to reappear on your machine, I'm afraid that we'll need to quirk it in the EC driver.

Thanks!

Quoted 16 lines · Rafael J. Wysocki

From: Rafael J. Wysocki <[email protected]>

Flushing the EC work while suspended to idle when the EC GPE status is not set causes some EC wakeup events (notably power button and lid ones) to be missed after a series of spurious wakeups on the Dell XPS13 9360 in my office.

If that happens, the machine cannot be woken up from suspend-to-idle by a power button press or lid status change and it needs to be woken up in some other way (eg. by a key press).

Flushing the EC work only after successful dispatching the EC GPE, which means that its status has been set, avoids the issue, so change the code in question accordingly.

Quoted 7 lines

Signed-off-by Rafael J. Wysocki <[email protected]>

---

Hi Chris,

Please check if the key press wakeup still works on your system with this patch applied (on top of https://patchwork.kernel.org/patch/11538065/).

Hi Chris,

Since I haven't heard back from you and the problem at hand is a regression on the machine where it happens, I'm going to push this patch for merging.

If it causes the key press wakeup issue to reappear on your machine, I'm afraid that we'll need to quirk it in the EC driver.

Thanks!

Quoted 75 lines

---
 drivers/acpi/ec.c    |    6 +++++-
 drivers/acpi/sleep.c |   15 ++++-----------
 2 files changed, 9 insertions(+), 12 deletions(-)

Index: linux-pm/drivers/acpi/ec.c
===================================================================

drivers/acpi/ec.c +5 -1

@@ -2020,9 +2020,13 @@ bool acpi_ec_dispatch_gpe(void)
         * to allow the caller to process events properly after that.
         */
        ret = acpi_dispatch_gpe(NULL, first_ec->gpe);
-       if (ret == ACPI_INTERRUPT_HANDLED)
+       if (ret == ACPI_INTERRUPT_HANDLED) {
                pm_pr_dbg("EC GPE dispatched\n");
 
+               /* Flush the event and query workqueues. */
+               acpi_ec_flush_work();
+       }
+
        return false;
 }
 #endif /* CONFIG_PM_SLEEP */

Index: linux-pm/drivers/acpi/sleep.c
===================================================================

drivers/acpi/sleep.c +4 -11

@@ -980,13 +980,6 @@ static int acpi_s2idle_prepare_late(void
        return 0;
 }
 
-static void acpi_s2idle_sync(void)
-{
-       /* The EC driver uses special workqueues that need to be flushed. */
-       acpi_ec_flush_work();
-       acpi_os_wait_events_complete(); /* synchronize Notify handling */
-}
-
 static bool acpi_s2idle_wake(void)
 {
        if (!acpi_sci_irq_valid())

@@ -1018,7 +1011,7 @@ static bool acpi_s2idle_wake(void)
                        return true;
 
                /*
-                * Cancel the wakeup and process all pending events in case
+                * Cancel the SCI wakeup and process all pending events in case
                 * there are any wakeup ones in there.
                 *
                 * Note that if any non-EC GPEs are active at this point, the

@@ -1026,8 +1019,7 @@ static bool acpi_s2idle_wake(void)
                 * should be missed by canceling the wakeup here.
                 */
                pm_system_cancel_wakeup();
-
-               acpi_s2idle_sync();
+               acpi_os_wait_events_complete();
 
                /*
                 * The SCI is in the "suspended" state now and it cannot produce

@@ -1060,7 +1052,8 @@ static void acpi_s2idle_restore(void)
         * of GPEs.
         */
        acpi_os_wait_events_complete(); /* synchronize GPE processing */
-       acpi_s2idle_sync();
+       acpi_ec_flush_work(); /* flush the EC driver's workqueues */
+       acpi_os_wait_events_complete(); /* synchronize Notify handling */
 
        s2idle_wakeup = false;
 

Quoted 32 lines · Rafael J. Wysocki

Quoted 16 lines · Rafael J. Wysocki

From: Rafael J. Wysocki <[email protected]>

Flushing the EC work while suspended to idle when the EC GPE status is not set causes some EC wakeup events (notably power button and lid ones) to be missed after a series of spurious wakeups on the Dell XPS13 9360 in my office.

If that happens, the machine cannot be woken up from suspend-to-idle by a power button press or lid status change and it needs to be woken up in some other way (eg. by a key press).

Flushing the EC work only after successful dispatching the EC GPE, which means that its status has been set, avoids the issue, so change the code in question accordingly.

Quoted 7 lines

Signed-off-by Rafael J. Wysocki <[email protected]>

---

Hi Chris,

Please check if the key press wakeup still works on your system with this patch applied (on top of https://patchwork.kernel.org/patch/11538065/).

Hi Chris,

Since I haven't heard back from you and the problem at hand is a regression on the machine where it happens, I'm going to push this patch for merging.

Quoted 4 lines

If it causes the key press wakeup issue to reappear on your machine, I'm afraid that we'll need to quirk it in the EC driver.

Thanks!

Hi Rafael,
My laptop works w/o problem waking up from a keystroke with this patch on top of https://patchwork.kernel.org/patch/11538065/).

Quoted 74 lines

Quoted 74 lines

---
 drivers/acpi/ec.c    |    6 +++++-
 drivers/acpi/sleep.c |   15 ++++-----------
 2 files changed, 9 insertions(+), 12 deletions(-)

Index: linux-pm/drivers/acpi/ec.c
===================================================================

drivers/acpi/ec.c +5 -1

@@ -2020,9 +2020,13 @@ bool acpi_ec_dispatch_gpe(void)
         * to allow the caller to process events properly after that.
         */
        ret = acpi_dispatch_gpe(NULL, first_ec->gpe);
-       if (ret == ACPI_INTERRUPT_HANDLED)
+       if (ret == ACPI_INTERRUPT_HANDLED) {
                pm_pr_dbg("EC GPE dispatched\n");
 
+               /* Flush the event and query workqueues. */
+               acpi_ec_flush_work();
+       }
+
        return false;
 }
 #endif /* CONFIG_PM_SLEEP */

Index: linux-pm/drivers/acpi/sleep.c
===================================================================

drivers/acpi/sleep.c +4 -11

@@ -980,13 +980,6 @@ static int acpi_s2idle_prepare_late(void
        return 0;
 }
 
-static void acpi_s2idle_sync(void)
-{
-       /* The EC driver uses special workqueues that need to be flushed. */
-       acpi_ec_flush_work();
-       acpi_os_wait_events_complete(); /* synchronize Notify handling */
-}
-
 static bool acpi_s2idle_wake(void)
 {
        if (!acpi_sci_irq_valid())

@@ -1018,7 +1011,7 @@ static bool acpi_s2idle_wake(void)
                        return true;
 
                /*
-                * Cancel the wakeup and process all pending events in case
+                * Cancel the SCI wakeup and process all pending events in case
                 * there are any wakeup ones in there.
                 *
                 * Note that if any non-EC GPEs are active at this point, the

@@ -1026,8 +1019,7 @@ static bool acpi_s2idle_wake(void)
                 * should be missed by canceling the wakeup here.
                 */
                pm_system_cancel_wakeup();
-
-               acpi_s2idle_sync();
+               acpi_os_wait_events_complete();
 
                /*
                 * The SCI is in the "suspended" state now and it cannot produce

@@ -1060,7 +1052,8 @@ static void acpi_s2idle_restore(void)
         * of GPEs.
         */
        acpi_os_wait_events_complete(); /* synchronize GPE processing */
-       acpi_s2idle_sync();
+       acpi_ec_flush_work(); /* flush the EC driver's workqueues */
+       acpi_os_wait_events_complete(); /* synchronize Notify handling */
 
        s2idle_wakeup = false;