Skip to contentSkip to conversation

eeepc-laptop: fix hot-unplug on resume

PATCH In review Alan Jenkins 3 participants 4 messages Jun 29, 2009

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 eeepc-laptop: fix hot-unplug on resume

The patch below fixes a regression introduced by the rfkill rewrite. Please consider it for inclusion in 2.6.31 via the ACPI tree.

Thanks
Alan

-------------------->

Quoted 1 line

From d0265f0bfe63086816d61c69ee4b2dc4cb491d06 Mon Sep 17 00:00:00 2001

From: Alan Jenkins <[email protected]>
Date: Mon, 29 Jun 2009 09:40:07 +0100
Subject: [PATCH] eeepc-laptop: fix hot-unplug on resume

OOPS on resume when the wireless adaptor is disabled during suspend was introduced by "eeepc-laptop: read rfkill soft-blocked state on resume".

Unable to handle kernel NULL pointer dereference

Process s2disk

Tainted: G W

IP: klist_put

Call trace:

? klist_del
? device_del
? device_unregister
? pci_stop_dev
? pci_stop_bus
? pci_remove_device
? eeepc_rfkill_hotplug [eeepc_laptop]
? eeepc_hotk_resume [eeepc_laptop]
? acpi_device_resume
? device_resume
? hibernation_snapshot

It appears the PCI device is removed twice. The eeepc_rfkill_hotplug() call from the resume handler is racing against the call from the ACPI notifier callback. The ACPI notification is triggered by the resume handler when it refreshes the value of CM_ASL_WLAN.

The fix is to serialize hotplug calls using a workqueue.

---
 drivers/platform/x86/eeepc-laptop.c |    9 ++++++---
 1 files changed, 6 insertions(+), 3 deletions(-)
drivers/platform/x86/eeepc-laptop.c +6 -3
index ec560f1..222ffb8 100644
@@ -143,6 +143,7 @@ struct eeepc_hotk {
 	struct rfkill *bluetooth_rfkill;
 	struct rfkill *wwan3g_rfkill;
 	struct hotplug_slot *hotplug_slot;
+	struct work_struct hotplug_work;
 };
 
 /* The actual device the driver binds to */
@@ -660,7 +661,7 @@ static int eeepc_get_adapter_status(struct hotplug_slot *hotplug_slot,
 	return 0;
 }
 
-static void eeepc_rfkill_hotplug(void)
+static void eeepc_hotplug_work(struct work_struct *work)
 {
 	struct pci_dev *dev;
 	struct pci_bus *bus = pci_find_bus(0, 1);
@@ -701,7 +702,7 @@ static void eeepc_rfkill_notify(acpi_handle handle, u32 event, void *data)
 	if (event != ACPI_NOTIFY_BUS_CHECK)
 		return;
 
-	eeepc_rfkill_hotplug();
+	schedule_work(&ehotk->hotplug_work);
 }
 
 static void eeepc_hotk_notify(struct acpi_device *device, u32 event)
@@ -892,7 +893,7 @@ static int eeepc_hotk_resume(struct acpi_device *device)
 
 		rfkill_set_sw_state(ehotk->wlan_rfkill, wlan != 1);
 
-		eeepc_rfkill_hotplug();
+		schedule_work(&ehotk->hotplug_work);
 	}
 
 	if (ehotk->bluetooth_rfkill)
@@ -1093,6 +1094,8 @@ static int eeepc_rfkill_init(struct device *dev)
 {
 	int result = 0;
 
+	INIT_WORK(&ehotk->hotplug_work, eeepc_hotplug_work);
+
 	eeepc_register_rfkill_notifier("\\_SB.PCI0.P0P6");
 	eeepc_register_rfkill_notifier("\\_SB.PCI0.P0P7");
 
Signature · 1 line
1.6.3.2
maintainer
#1 lore raw
Quoted 5 lines · Alan Jenkins

The patch below fixes a regression introduced by the rfkill rewrite. Please consider it for inclusion in 2.6.31 via the ACPI tree.

Thanks
Alan

Hi Len,

Please merge this patch in ACPI tree.
Thanks,

Signature · 2 lines
Corentin Chary
http://xf.iksaif.net - http://uffs.org
#2 lore raw
In reply to Alan Jenkins
Quoted 2 lines · Alan Jenkins

The patch below fixes a regression introduced by the rfkill rewrite. Please consider it for inclusion in 2.6.31 via the ACPI tree.

Is there a Bugzilla # for this regression?

Rafael

Quoted 88 lines

-------------------->

From d0265f0bfe63086816d61c69ee4b2dc4cb491d06 Mon Sep 17 00:00:00 2001
From: Alan Jenkins <[email protected]>
Date: Mon, 29 Jun 2009 09:40:07 +0100
Subject: [PATCH] eeepc-laptop: fix hot-unplug on resume

OOPS on resume when the wireless adaptor is disabled during suspend was introduced by "eeepc-laptop: read rfkill soft-blocked state on resume".

Unable to handle kernel NULL pointer dereference

Process s2disk

Tainted: G W

IP: klist_put

Call trace:

? klist_del
? device_del
? device_unregister
? pci_stop_dev
? pci_stop_bus
? pci_remove_device
? eeepc_rfkill_hotplug [eeepc_laptop]
? eeepc_hotk_resume [eeepc_laptop]
? acpi_device_resume
? device_resume
? hibernation_snapshot

It appears the PCI device is removed twice. The eeepc_rfkill_hotplug() call from the resume handler is racing against the call from the ACPI notifier callback. The ACPI notification is triggered by the resume handler when it refreshes the value of CM_ASL_WLAN.

The fix is to serialize hotplug calls using a workqueue.

---
 drivers/platform/x86/eeepc-laptop.c |    9 ++++++---
 1 files changed, 6 insertions(+), 3 deletions(-)
drivers/platform/x86/eeepc-laptop.c +6 -3
index ec560f1..222ffb8 100644
@@ -143,6 +143,7 @@ struct eeepc_hotk {
 	struct rfkill *bluetooth_rfkill;
 	struct rfkill *wwan3g_rfkill;
 	struct hotplug_slot *hotplug_slot;
+	struct work_struct hotplug_work;
 };
 
 /* The actual device the driver binds to */
@@ -660,7 +661,7 @@ static int eeepc_get_adapter_status(struct hotplug_slot *hotplug_slot,
 	return 0;
 }
 
-static void eeepc_rfkill_hotplug(void)
+static void eeepc_hotplug_work(struct work_struct *work)
 {
 	struct pci_dev *dev;
 	struct pci_bus *bus = pci_find_bus(0, 1);
@@ -701,7 +702,7 @@ static void eeepc_rfkill_notify(acpi_handle handle, u32 event, void *data)
 	if (event != ACPI_NOTIFY_BUS_CHECK)
 		return;
 
-	eeepc_rfkill_hotplug();
+	schedule_work(&ehotk->hotplug_work);
 }
 
 static void eeepc_hotk_notify(struct acpi_device *device, u32 event)
@@ -892,7 +893,7 @@ static int eeepc_hotk_resume(struct acpi_device *device)
 
 		rfkill_set_sw_state(ehotk->wlan_rfkill, wlan != 1);
 
-		eeepc_rfkill_hotplug();
+		schedule_work(&ehotk->hotplug_work);
 	}
 
 	if (ehotk->bluetooth_rfkill)
@@ -1093,6 +1094,8 @@ static int eeepc_rfkill_init(struct device *dev)
 {
 	int result = 0;
 
+	INIT_WORK(&ehotk->hotplug_work, eeepc_hotplug_work);
+
 	eeepc_register_rfkill_notifier("\\_SB.PCI0.P0P6");
 	eeepc_register_rfkill_notifier("\\_SB.PCI0.P0P7");
#3 lore raw
Quoted 10 lines · Rafael J. Wysocki
Quoted 3 lines · Alan Jenkins

The patch below fixes a regression introduced by the rfkill rewrite. Please consider it for inclusion in 2.6.31 via the ACPI tree.

Is there a Bugzilla # for this regression?

Rafael

No.

I found it in testing, a few days after I'd submitted the change myself :-/.

Thanks
Alan