Skip to contentSkip to conversation

ACPI / PM: Fix incorrect wakeup irq setting before suspend-to-idle

PATCH In review Chen Yu 3 participants 6 messages Oct 9, 2015 – Oct 14, 2015 ACPI

lore
Share & tools
lore

Chen Yu

author
#0 lore raw

PATCH v3 1/1 ACPI / PM: Fix incorrect wakeup irq setting before suspend-to-idle

For ACPI compatible system, SCI(ACPI System Control Interrupt) is used to wake system up from suspend-to-idle. Once CPU is woken up by SCI, interrupt handler will firstly checks if current interrupt is legal to wake up the whole system, thus irq_pm_check_wakeup is invoked to validate the irq number. However, before suspend-to-idle, acpi_gbl_FADT.sci_interrupt is marked rather than actual irq number in acpi_freeze_prepare, this might lead to unable to wake up the system.

This patch fixes this problem by marking the irq number return by acpi_gsi_to_irq as IRQD_WAKEUP_STATE, rather than marking the acpi_gbl_FADT.sci_interrupt. Meanwhile this patch fixes the same problems inside acpi_os_remove_interrupt_handler and acpi_os_wait_events_complete respectively.

---
v3:
- 1.Rename acpi_inuse_irq to acpi_sci_irq for better understanding. 2.If the irq handler is not registered, skip the synchronize_hardirq in acpi_os_wait_events_complete and return immediately in acpi_os_remove_interrupt_handler.
3.For acpi_freeze_prepare and acpi_freeze_restore, if the acpi irq
handler is not properly registered, we do not leverage acpi irq to wake up the system, but expect other peripherals(such as PCI devices and USB devices)to invoke enable_irq_wake for us.
v2:
- 1.Define a global acpi_inuse_irq variable, store irq in it and access it directly from acpi_freeze_prepare(), and it doesn't have to depend on CONFIG_SUSPEND as it is just the IRQ number actually used by ACPI.
2.Also fix the same problems inside acpi_os_remove_interrupt_handler,
acpi_os_wait_events_complete.

---
 drivers/acpi/osl.c   | 12 ++++++++----
 drivers/acpi/sleep.c |  6 ++++--
 include/linux/acpi.h |  3 +++
 3 files changed, 15 insertions(+), 6 deletions(-)
drivers/acpi/osl.c +8 -4
index 739a4a6..8e1a5d3 100644
@@ -81,6 +81,7 @@ static struct workqueue_struct *kacpid_wq;
 static struct workqueue_struct *kacpi_notify_wq;
 static struct workqueue_struct *kacpi_hotplug_wq;
 static bool acpi_os_initialized;
+unsigned int acpi_sci_irq = INVALID_ACPI_IRQ;
 
 /*
  * This list of permanent mappings is for memory that may be accessed from
@@ -856,17 +857,20 @@ acpi_os_install_interrupt_handler(u32 gsi, acpi_osd_handler handler,
 		acpi_irq_handler = NULL;
 		return AE_NOT_ACQUIRED;
 	}
+	acpi_sci_irq = irq;
 
 	return AE_OK;
 }
 
 acpi_status acpi_os_remove_interrupt_handler(u32 irq, acpi_osd_handler handler)
 {
-	if (irq != acpi_gbl_FADT.sci_interrupt)
+	if ((irq != acpi_gbl_FADT.sci_interrupt) ||
+			IS_INVALID_ACPI_IRQ(acpi_sci_irq))
 		return AE_BAD_PARAMETER;
 
-	free_irq(irq, acpi_irq);
+	free_irq(acpi_sci_irq, acpi_irq);
 	acpi_irq_handler = NULL;
+	acpi_sci_irq = INVALID_ACPI_IRQ;
 
 	return AE_OK;
 }
@@ -1180,8 +1184,8 @@ void acpi_os_wait_events_complete(void)
 	 * Make sure the GPE handler or the fixed event handler is not used
 	 * on another CPU after removal.
 	 */
-	if (acpi_irq_handler)
-		synchronize_hardirq(acpi_gbl_FADT.sci_interrupt);
+	if (!IS_INVALID_ACPI_IRQ(acpi_sci_irq))
+		synchronize_hardirq(acpi_sci_irq);
 	flush_workqueue(kacpid_wq);
 	flush_workqueue(kacpi_notify_wq);
 }
drivers/acpi/sleep.c +4 -2
index 2f0d4db..1704595 100644
@@ -632,14 +632,16 @@ static int acpi_freeze_prepare(void)
 	acpi_enable_wakeup_devices(ACPI_STATE_S0);
 	acpi_enable_all_wakeup_gpes();
 	acpi_os_wait_events_complete();
-	enable_irq_wake(acpi_gbl_FADT.sci_interrupt);
+	if (!IS_INVALID_ACPI_IRQ(acpi_sci_irq))
+		enable_irq_wake(acpi_sci_irq);
 	return 0;
 }
 
 static void acpi_freeze_restore(void)
 {
 	acpi_disable_wakeup_devices(ACPI_STATE_S0);
-	disable_irq_wake(acpi_gbl_FADT.sci_interrupt);
+	if (!IS_INVALID_ACPI_IRQ(acpi_sci_irq))
+		disable_irq_wake(acpi_sci_irq);
 	acpi_enable_all_runtime_gpes();
 }
 
include/linux/acpi.h +3 -0
index 43856d1..2f05dc0 100644
@@ -193,6 +193,9 @@ int acpi_ioapic_registered(acpi_handle handle, u32 gsi_base);
 void acpi_irq_stats_init(void);
 extern u32 acpi_irq_handled;
 extern u32 acpi_irq_not_handled;
+extern unsigned int acpi_sci_irq;
+#define INVALID_ACPI_IRQ  ((unsigned)-1)
+#define IS_INVALID_ACPI_IRQ(x) unlikely((x) == INVALID_ACPI_IRQ)
 
 extern int sbf_port;
 extern unsigned long acpi_realmode_flags;
Signature · 1 line
1.8.4.2
#1 lore raw

Hi, Yu

Quoted 56 lines

From: Chen, Yu C
Sent: Friday, October 09, 2015 4:20 PM

For ACPI compatible system, SCI(ACPI System Control Interrupt) is used to wake system up from suspend-to-idle. Once CPU is woken up by SCI, interrupt handler will firstly checks if current interrupt is legal to wake up the whole system, thus irq_pm_check_wakeup is invoked to validate the irq number. However, before suspend-to-idle, acpi_gbl_FADT.sci_interrupt is marked rather than actual irq number in acpi_freeze_prepare, this might lead to unable to wake up the system.

This patch fixes this problem by marking the irq number return by acpi_gsi_to_irq as IRQD_WAKEUP_STATE, rather than marking the acpi_gbl_FADT.sci_interrupt. Meanwhile this patch fixes the same problems inside acpi_os_remove_interrupt_handler and acpi_os_wait_events_complete respectively.

---
v3:
- 1.Rename acpi_inuse_irq to acpi_sci_irq for better understanding. 2.If the irq handler is not registered, skip the synchronize_hardirq in acpi_os_wait_events_complete and return immediately in acpi_os_remove_interrupt_handler.
3.For acpi_freeze_prepare and acpi_freeze_restore, if the acpi irq
handler is not properly registered, we do not leverage acpi irq to wake up the system, but expect other peripherals(such as PCI devices and USB devices)to invoke enable_irq_wake for us.
v2:
- 1.Define a global acpi_inuse_irq variable, store irq in it and access it directly from acpi_freeze_prepare(), and it doesn't have to depend on CONFIG_SUSPEND as it is just the IRQ number actually used by ACPI.
2.Also fix the same problems inside acpi_os_remove_interrupt_handler,
acpi_os_wait_events_complete.

---
 drivers/acpi/osl.c   | 12 ++++++++----
 drivers/acpi/sleep.c |  6 ++++--
 include/linux/acpi.h |  3 +++
 3 files changed, 15 insertions(+), 6 deletions(-)
drivers/acpi/osl.c +1 -0
index 739a4a6..8e1a5d3 100644
@@ -81,6 +81,7 @@ static struct workqueue_struct *kacpid_wq;
 static struct workqueue_struct *kacpi_notify_wq;
 static struct workqueue_struct *kacpi_hotplug_wq;
 static bool acpi_os_initialized;
+unsigned int acpi_sci_irq = INVALID_ACPI_IRQ;
 
 /*
  * This list of permanent mappings is for memory that may be accessed from

drivers/acpi/osl.c acpi_os_install_interrupt_handler(u32 gsi, acpi_osd_handler handler, · line 865

Quoted 9 lines
drivers/acpi/osl.c +1 -0
@@ -856,17 +857,20 @@ acpi_os_install_interrupt_handler(u32 gsi, acpi_osd_handler handler,
 		acpi_irq_handler = NULL;
 		return AE_NOT_ACQUIRED;
 	}
+	acpi_sci_irq = irq;
 
 	return AE_OK;
 }
 
 acpi_status acpi_os_remove_interrupt_handler(u32 irq, acpi_osd_handler handler)

Why don't you rename irq here to gsi to improve the readability?
The false naming and the wrong example written for this function are probably the root causes of all other bad code. So if we want to stop people making future mistakes, we need to cleanup ourselves.

Thanks and best regards
-Lv

Quoted 63 lines
drivers/acpi/osl.c +6 -4
 {
-	if (irq != acpi_gbl_FADT.sci_interrupt)
+	if ((irq != acpi_gbl_FADT.sci_interrupt) ||
+			IS_INVALID_ACPI_IRQ(acpi_sci_irq))
 		return AE_BAD_PARAMETER;
 
-	free_irq(irq, acpi_irq);
+	free_irq(acpi_sci_irq, acpi_irq);
 	acpi_irq_handler = NULL;
+	acpi_sci_irq = INVALID_ACPI_IRQ;
 
 	return AE_OK;
 }
@@ -1180,8 +1184,8 @@ void acpi_os_wait_events_complete(void)
 	 * Make sure the GPE handler or the fixed event handler is not used
 	 * on another CPU after removal.
 	 */
-	if (acpi_irq_handler)
-		synchronize_hardirq(acpi_gbl_FADT.sci_interrupt);
+	if (!IS_INVALID_ACPI_IRQ(acpi_sci_irq))
+		synchronize_hardirq(acpi_sci_irq);
 	flush_workqueue(kacpid_wq);
 	flush_workqueue(kacpi_notify_wq);
 }
drivers/acpi/sleep.c +4 -2
index 2f0d4db..1704595 100644
@@ -632,14 +632,16 @@ static int acpi_freeze_prepare(void)
 	acpi_enable_wakeup_devices(ACPI_STATE_S0);
 	acpi_enable_all_wakeup_gpes();
 	acpi_os_wait_events_complete();
-	enable_irq_wake(acpi_gbl_FADT.sci_interrupt);
+	if (!IS_INVALID_ACPI_IRQ(acpi_sci_irq))
+		enable_irq_wake(acpi_sci_irq);
 	return 0;
 }
 
 static void acpi_freeze_restore(void)
 {
 	acpi_disable_wakeup_devices(ACPI_STATE_S0);
-	disable_irq_wake(acpi_gbl_FADT.sci_interrupt);
+	if (!IS_INVALID_ACPI_IRQ(acpi_sci_irq))
+		disable_irq_wake(acpi_sci_irq);
 	acpi_enable_all_runtime_gpes();
 }
 
include/linux/acpi.h +3 -1
index 43856d1..2f05dc0 100644
@@ -193,6 +193,9 @@ int acpi_ioapic_registered(acpi_handle handle, u32 gsi_base);
 void acpi_irq_stats_init(void);
 extern u32 acpi_irq_handled;
 extern u32 acpi_irq_not_handled;
+extern unsigned int acpi_sci_irq;
+#define INVALID_ACPI_IRQ  ((unsigned)-1)
+#define IS_INVALID_ACPI_IRQ(x) unlikely((x) == INVALID_ACPI_IRQ)
 
 extern int sbf_port;
 extern unsigned long acpi_realmode_flags;
--

1.8.4.2

author
#2 lore raw

Hi, LV

Quoted 15 lines

-----Original Message-----
From: Zheng, Lv
Sent: Friday, October 09, 2015 4:33 PM
To: Chen, Yu C; [email protected]; [email protected]
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Rui
Subject: RE: [PATCH][v3] ACPI / PM: Fix incorrect wakeup irq setting before suspend-to-idle

Hi, Yu

Quoted 4 lines

From: Chen, Yu C
Sent: Friday, October 09, 2015 4:20 PM

Quoted 9 lines
Quoted 2 lines

acpi_status acpi_os_remove_interrupt_handler(u32 irq,
acpi_osd_handler handler)

Why don't you rename irq here to gsi to improve the readability? The false naming and the wrong example written for this function are probably the root causes of all other bad code.
So if we want to stop people making future mistakes, we need to cleanup ourselves.

OK, will rewrite in next version.

Quoted 3 lines

Thanks and best regards
-Lv

#3 lore raw

Hi,

Quoted 17 lines

From: Chen, Yu C
Sent: Friday, October 09, 2015 5:50 PM

Hi, LV

Quoted 12 lines

From: Zheng, Lv
Sent: Friday, October 09, 2015 4:33 PM

Hi, Yu

Quoted 6 lines

From: Chen, Yu C
Sent: Friday, October 09, 2015 4:20 PM

acpi_status acpi_os_remove_interrupt_handler(u32 irq,
acpi_osd_handler handler)

Quoted 7 lines
Quoted 6 lines

Why don't you rename irq here to gsi to improve the readability? The false naming and the wrong example written for this function are probably the root causes of all other bad code.
So if we want to stop people making future mistakes, we need to cleanup ourselves.

OK, will rewrite in next version.

You can add Acked-by: Lv Zheng <[email protected]> And don't forget to mark it as a stable material.

One more question is:
Do you want the modules to use "acpi_sci_irq"? If the answer is no, then please ignore this question.

Thanks and best regards
-Lv

Quoted 4 lines
Quoted 4 lines

Thanks and best regards
-Lv

#4 lore raw
In reply to Chen, Yu C
Quoted 20 lines · Yu C

Hi, LV

Quoted 18 lines

-----Original Message-----
From: Zheng, Lv
Sent: Friday, October 09, 2015 4:33 PM
To: Chen, Yu C; [email protected]; [email protected]
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Rui
Subject: RE: [PATCH][v3] ACPI / PM: Fix incorrect wakeup irq setting before suspend-to-idle

Hi, Yu

Quoted 6 lines

From: Chen, Yu C
Sent: Friday, October 09, 2015 4:20 PM

acpi_status acpi_os_remove_interrupt_handler(u32 irq,
acpi_osd_handler handler)

Quoted 7 lines
Quoted 6 lines

Why don't you rename irq here to gsi to improve the readability? The false naming and the wrong example written for this function are probably the root causes of all other bad code.
So if we want to stop people making future mistakes, we need to cleanup ourselves.

OK, will rewrite in next version.

It would be good to change the subject of the patch too I think, because it is not about wakeup IRQs only any more.

Thanks,
Rafael

author
#5 lore raw

Hi ,Rafael,

Quoted 36 lines

-----Original Message-----
From: Rafael J. Wysocki [mailto:[email protected]]
Sent: Tuesday, October 13, 2015 4:20 AM
To: Chen, Yu C
Cc: Zheng, Lv; [email protected]; [email protected]; linux- [email protected]; Zhang, Rui; Wysocki, Rafael J; Brown, Len
Subject: Re: [PATCH][v3] ACPI / PM: Fix incorrect wakeup irq setting before suspend-to-idle

Quoted 26 lines · Yu C

Hi, LV

Quoted 24 lines

-----Original Message-----
From: Zheng, Lv
Sent: Friday, October 09, 2015 4:33 PM
To: Chen, Yu C; [email protected]; [email protected]
Cc: [email protected]; [email protected]; linux- [email protected]; Zhang, Rui
Subject: RE: [PATCH][v3] ACPI / PM: Fix incorrect wakeup irq setting before suspend-to-idle

Hi, Yu

Quoted 6 lines

From: Chen, Yu C
Sent: Friday, October 09, 2015 4:20 PM

acpi_status acpi_os_remove_interrupt_handler(u32 irq,
acpi_osd_handler handler)

Why don't you rename irq here to gsi to improve the readability? The false naming and the wrong example written for this function are probably the root causes of all other bad code.
So if we want to stop people making future mistakes, we need to cleanup ourselves.

Quoted 5 lines
Quoted 1 line

OK, will rewrite in next version.

It would be good to change the subject of the patch too I think, because it is not about wakeup IRQs only any more.

OK, will send another series of patches.

Best Regards,
Yu