ACPI / WAKEUP : enable wakeup power for physical child devices

commit f2b56bc808addb908a5bf435d9b942c02af9a7c4
("ACPI / PM: Use device wakeup flags for handling ACPI wakeup devices")

broke wake-on-lan for Broadcom BCM4401 Ethernet card (b44) on Dell Vostro 1500 laptop. device_may_wakeup for main ACPI device (PCIE) returns false, because it has can_wakeup = 0. Therefore any physical devices connected to this parent were not prepared for wakeup/suspend. But physical device is capable to wakeup the system.

To fix this issue device_may_wakeup was replaced with acpi_device_may_wakeup. acpi_device_may_wakeup was written based on function acpi_system_wakeup_device_seq_show

Signed-off-by Andrey Skvortsov <[email protected]>

---
 drivers/acpi/wakeup.c |   45 +++++++++++++++++++++++++++++++++++++--------
 1 files changed, 37 insertions(+), 8 deletions(-)

drivers/acpi/wakeup.c +28 -0

index 1638401..0da7e70 100644
@@ -19,6 +19,34 @@
 #define _COMPONENT		ACPI_SYSTEM_COMPONENT
 ACPI_MODULE_NAME("wakeup_devices")
 
+
+bool acpi_device_may_wakeup(struct acpi_device *dev)
+{
+	struct acpi_device_physical_node *entry;
+	bool may_wakeup = false;
+
+	mutex_lock(&dev->physical_node_lock);
+	if (!dev->physical_node_count)
+		may_wakeup = device_may_wakeup(&dev->dev);
+	else {
+		struct device *ldev;
+
+		list_for_each_entry(entry, &dev->physical_node_list, node) {
+			ldev = get_device(entry->dev);
+			if (!ldev)
+				continue;
+
+			may_wakeup = device_may_wakeup(&dev->dev) ||
+				device_may_wakeup(ldev);
+
+			put_device(ldev);
+		}
+	}
+
+	mutex_unlock(&dev->physical_node_lock);
+	return may_wakeup;
+}
+
 /**
  * acpi_enable_wakeup_devices - Enable wake-up device GPEs.
  * @sleep_state: ACPI system sleep state.

@@ -35,13 +63,14 @@ void acpi_enable_wakeup_devices(u8 sleep_state)
 		struct acpi_device *dev =

drivers/acpi/wakeup.c +4 -4

 			continue;
 
-		if (device_may_wakeup(&dev->dev))
+		if (acpi_device_may_wakeup(dev))
 			acpi_enable_wakeup_device_power(dev, sleep_state);
 
 		/* The wake-up power should have been enabled already. */

@@ -63,15 +92,15 @@ void acpi_disable_wakeup_devices(u8 sleep_state)
 			container_of(node, struct acpi_device, wakeup_list);
 
 		if (!dev->wakeup.flags.valid
-		    || sleep_state > (u32) dev->wakeup.sleep_state
-		    || !(device_may_wakeup(&dev->dev)
-		        || dev->wakeup.prepare_count))
+			|| sleep_state > (u32) dev->wakeup.sleep_state
+			|| !(acpi_device_may_wakeup(dev)
+				|| dev->wakeup.prepare_count))

Quoted 6 lines

Quoted 2 lines

-> dev->wakeup.prepare_count here shold be different from 0 at this point.

Thank you very much for the tip. I checked prepare_count, it was 0.
I tried fix that and patched ssb and b44 drivers, that are used for my Ethernet card. The code is below. Is this a better way to solve this problem?

Having a cursory look at the changes only, I think so.

Quoted 7 lines

drivers/net/ethernet/broadcom/b44.c +0 -0

index 416620f..ebac411 100644
@@ -2103,7 +2103,9 @@ static int b44_set_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
 	else
 		bp->flags &= ~B44_FLAG_WOL_ENABLE;

Wake On Lan was not working on laptop DELL Vostro 1500.
If WOL was turned on, BCM4401 was powered up in suspend mode. LEDs blinked. But the laptop could not be woken up with the Magic Packet. The reason for that was that PCIE was not enabled as a system wakeup source and therefore the host PCI bridge was not powered up in suspend mode. PCIE was not enabled in suspend by PM because no child devices were registered as wakeup source during suspend process.

On laptop BCM4401 is connected through the SSB bus, that is connected to the PCI-Express bus. SSB and B44 did not use standard PM wakeup functions
and did not forward wakeup settings to their parents.
To fix that B44 driver enables PM wakeup and registers new wakeup source using device_set_wakeup_enable(). Wakeup is automatically reported to the parent SSB bus via power.wakeup_path. SSB bus enables wakeup for the parent PCI bridge, if there is any child devices with enabled wakeup functionality. All other steps are
done by PM core code.

Signed-off-by Andrey Skvortsov <[email protected]>

---
 drivers/net/ethernet/broadcom/b44.c |    2 ++
 drivers/ssb/pcihost_wrapper.c       |   33 ++++++++++++++++++++++-----------
 2 files changed, 24 insertions(+), 11 deletions(-)

drivers/net/ethernet/broadcom/b44.c +2 -0

index 416620f..ffeaf47 100644
@@ -2104,6 +2104,7 @@ static int b44_set_wol(struct net_device *dev, struct ethtool_wolinfo *wol)
 		bp->flags &= ~B44_FLAG_WOL_ENABLE;
 	spin_unlock_irq(&bp->lock);
 
+	device_set_wakeup_enable(bp->sdev->dev, wol->wolopts & WAKE_MAGIC);
 	return 0;
 }
 

@@ -2452,6 +2453,7 @@ static int b44_init_one(struct ssb_device *sdev,
 		}
 	}
 
+	device_set_wakeup_capable(sdev->dev, true);
 	netdev_info(dev, "%s %pM\n", DRV_DESCRIPTION, dev->dev_addr);
 
 	return 0;

drivers/ssb/pcihost_wrapper.c +22 -11

index 69161bb..410215c 100644
@@ -11,15 +11,17 @@
  * Licensed under the GNU/GPL. See COPYING for details.
  */
 
+#include <linux/pm.h>
 #include <linux/pci.h>
 #include <linux/export.h>
 #include <linux/slab.h>
 #include <linux/ssb/ssb.h>
 
 
-#ifdef CONFIG_PM
-static int ssb_pcihost_suspend(struct pci_dev *dev, pm_message_t state)
+#ifdef CONFIG_PM_SLEEP
+static int ssb_pcihost_suspend(struct device *d)
 {
+	struct pci_dev *dev = to_pci_dev(d);
 	struct ssb_bus *ssb = pci_get_drvdata(dev);
 	int err;
 

@@ -28,17 +30,23 @@ static int ssb_pcihost_suspend(struct pci_dev *dev, pm_message_t state)
 		return err;
 	pci_save_state(dev);
 	pci_disable_device(dev);
-	pci_set_power_state(dev, pci_choose_state(dev, state));
+
+	/* if there is a wakeup enabled child device on ssb bus,
+	   enable pci wakeup posibility. */
+	device_set_wakeup_enable(d, d->power.wakeup_path);
+
+	pci_prepare_to_sleep(dev);
 
 	return 0;
 }
 
-static int ssb_pcihost_resume(struct pci_dev *dev)
+static int ssb_pcihost_resume(struct device *d)
 {
+	struct pci_dev *dev = to_pci_dev(d);
 	struct ssb_bus *ssb = pci_get_drvdata(dev);
 	int err;
 
-	pci_set_power_state(dev, PCI_D0);
+	pci_back_from_sleep(dev);
 	err = pci_enable_device(dev);
 	if (err)
 		return err;

@@ -49,10 +57,12 @@ static int ssb_pcihost_resume(struct pci_dev *dev)
 
 	return 0;
 }
-#else /* CONFIG_PM */
-# define ssb_pcihost_suspend	NULL
-# define ssb_pcihost_resume	NULL
-#endif /* CONFIG_PM */
+
+static const struct dev_pm_ops ssb_pcihost_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(ssb_pcihost_suspend, ssb_pcihost_resume)
+};
+
+#endif /* CONFIG_PM_SLEEP */
 
 static int ssb_pcihost_probe(struct pci_dev *dev,
 			     const struct pci_device_id *id)

@@ -115,8 +125,9 @@ int ssb_pcihost_register(struct pci_driver *driver)
 {
 	driver->probe = ssb_pcihost_probe;
 	driver->remove = ssb_pcihost_remove;
-	driver->suspend = ssb_pcihost_suspend;
-	driver->resume = ssb_pcihost_resume;
+#ifdef CONFIG_PM_SLEEP
+	driver->driver.pm = &ssb_pcihost_pm_ops;
+#endif
 
 	return pci_register_driver(driver);
 }

Quoted 7 lines · On Mon, 1 Dec 2014 23:46:38 +0300 Andrey Skvortsov…

Wake On Lan was not working on laptop DELL Vostro 1500.
If WOL was turned on, BCM4401 was powered up in suspend mode. LEDs blinked. But the laptop could not be woken up with the Magic Packet. The reason for that was that PCIE was not enabled as a system wakeup source and therefore the host PCI bridge was not powered up in suspend mode. PCIE was not enabled in suspend by PM because no child devices were registered as wakeup source during suspend process.

Quoted 8 lines

On laptop BCM4401 is connected through the SSB bus, that is connected to the PCI-Express bus. SSB and B44 did not use standard PM wakeup functions
and did not forward wakeup settings to their parents.
To fix that B44 driver enables PM wakeup and registers new wakeup source using device_set_wakeup_enable(). Wakeup is automatically reported to the parent SSB bus via power.wakeup_path. SSB bus enables wakeup for the parent PCI bridge, if there is any child devices with enabled wakeup functionality. All other steps are
done by PM core code.

Thanks, this looks good.
I assume you tested this (I currently don't have a device to test this).

Larry, Rafa?, any other b43 user:
Can you please test whether this doesn't cause regressions for suspend/resume on b43? (Patch is attached as reference)

Quoted 5 lines

Signed-off-by Andrey Skvortsov <[email protected]>

---
 drivers/net/ethernet/broadcom/b44.c |    2 ++
 drivers/ssb/pcihost_wrapper.c       |   33 ++++++++++++++++++++++-----------
 2 files changed, 24 insertions(+), 11 deletions(-)

Quoted 19 lines · Michael B?sch

Quoted 15 lines · On Mon, 1 Dec 2014 23:46:38 +0300 Andrey Skvortsov…

Wake On Lan was not working on laptop DELL Vostro 1500.
If WOL was turned on, BCM4401 was powered up in suspend mode. LEDs blinked. But the laptop could not be woken up with the Magic Packet. The reason for that was that PCIE was not enabled as a system wakeup source and
therefore the host PCI bridge was not powered up in suspend mode.
PCIE was not enabled in suspend by PM because no child devices were registered as wakeup source during suspend process.
On laptop BCM4401 is connected through the SSB bus, that is connected to the PCI-Express bus. SSB and B44 did not use standard PM wakeup functions
and did not forward wakeup settings to their parents.
To fix that B44 driver enables PM wakeup and registers new wakeup source using device_set_wakeup_enable(). Wakeup is automatically reported to the parent SSB bus via power.wakeup_path. SSB bus enables wakeup for the parent PCI bridge, if there is any child devices with enabled wakeup functionality. All other steps are
done by PM core code.

Quoted 2 lines

Thanks, this looks good.
I assume you tested this (I currently don't have a device to test this).

Sure, I've tested it. WOL from suspend is working and after resume from hibernate Ethernet is working too.

Quoted 20 lines · On Tue, 2 Dec 2014 23:01:29 +0300 Andrey Skvortsov…

Quoted 19 lines · Michael B?sch

Quoted 15 lines · On Mon, 1 Dec 2014 23:46:38 +0300 Andrey Skvortsov…

Wake On Lan was not working on laptop DELL Vostro 1500.
If WOL was turned on, BCM4401 was powered up in suspend mode. LEDs blinked. But the laptop could not be woken up with the Magic Packet. The reason for that was that PCIE was not enabled as a system wakeup source and
therefore the host PCI bridge was not powered up in suspend mode.
PCIE was not enabled in suspend by PM because no child devices were registered as wakeup source during suspend process.
On laptop BCM4401 is connected through the SSB bus, that is connected to the PCI-Express bus. SSB and B44 did not use standard PM wakeup functions
and did not forward wakeup settings to their parents.
To fix that B44 driver enables PM wakeup and registers new wakeup source using device_set_wakeup_enable(). Wakeup is automatically reported to the parent SSB bus via power.wakeup_path. SSB bus enables wakeup for the parent PCI bridge, if there is any child devices with enabled wakeup functionality. All other steps are
done by PM core code.

Quoted 4 lines

Quoted 2 lines

Thanks, this looks good.
I assume you tested this (I currently don't have a device to test this).

Sure, I've tested it. WOL from suspend is working and after resume from hibernate Ethernet is working too.

That sounds good, indeed.
I'd still prefer, if someone with b43 (wireless) would test it, too.

Quoted 28 lines · Michael B?sch

Quoted 24 lines · On Tue, 2 Dec 2014 23:01:29 +0300 Andrey Skvortsov…

Quoted 19 lines · Michael B?sch

Quoted 7 lines · On Mon, 1 Dec 2014 23:46:38 +0300 Andrey Skvortsov…

Wake On Lan was not working on laptop DELL Vostro 1500.
If WOL was turned on, BCM4401 was powered up in suspend mode. LEDs blinked. But the laptop could not be woken up with the Magic Packet. The reason for that was that PCIE was not enabled as a system wakeup source and therefore the host PCI bridge was not powered up in suspend mode. PCIE was not enabled in suspend by PM because no child devices were registered as wakeup source during suspend process.

Quoted 8 lines

On laptop BCM4401 is connected through the SSB bus, that is connected to the PCI-Express bus. SSB and B44 did not use standard PM wakeup functions
and did not forward wakeup settings to their parents.
To fix that B44 driver enables PM wakeup and registers new wakeup source using device_set_wakeup_enable(). Wakeup is automatically reported to the parent SSB bus via power.wakeup_path. SSB bus enables wakeup for the parent PCI bridge, if there is any child devices with enabled wakeup functionality. All other steps are
done by PM core code.

Quoted 2 lines

Thanks, this looks good.
I assume you tested this (I currently don't have a device to test this).

Sure, I've tested it. WOL from suspend is working and after resume from hibernate Ethernet is working too.

Quoted 2 lines

That sounds good, indeed.
I'd still prefer, if someone with b43 (wireless) would test it, too.

I did a partial test with my PowerBook G4. With the patch installed, it would both suspend and hibernate, but WOL would be impossible. This computer uses a PCMCIA version of the BCM4318, and power is turned off to the PCMCIA card when suspended or hibernating.

Quoted 29 lines · On Tue, 02 Dec 2014 16:23:49 -0600 Larry Finger <[email protected]>…

Quoted 28 lines · Michael B?sch

Quoted 24 lines · On Tue, 2 Dec 2014 23:01:29 +0300 Andrey Skvortsov…

Quoted 19 lines · Michael B?sch

Quoted 7 lines · On Mon, 1 Dec 2014 23:46:38 +0300 Andrey Skvortsov…

Wake On Lan was not working on laptop DELL Vostro 1500.
If WOL was turned on, BCM4401 was powered up in suspend mode. LEDs blinked. But the laptop could not be woken up with the Magic Packet. The reason for that was that PCIE was not enabled as a system wakeup source and therefore the host PCI bridge was not powered up in suspend mode. PCIE was not enabled in suspend by PM because no child devices were registered as wakeup source during suspend process.

Quoted 8 lines

On laptop BCM4401 is connected through the SSB bus, that is connected to the PCI-Express bus. SSB and B44 did not use standard PM wakeup functions
and did not forward wakeup settings to their parents.
To fix that B44 driver enables PM wakeup and registers new wakeup source using device_set_wakeup_enable(). Wakeup is automatically reported to the parent SSB bus via power.wakeup_path. SSB bus enables wakeup for the parent PCI bridge, if there is any child devices with enabled wakeup functionality. All other steps are
done by PM core code.

Quoted 2 lines

Thanks, this looks good.
I assume you tested this (I currently don't have a device to test this).

Sure, I've tested it. WOL from suspend is working and after resume from hibernate Ethernet is working too.

Quoted 7 lines

Quoted 2 lines

That sounds good, indeed.
I'd still prefer, if someone with b43 (wireless) would test it, too.

I did a partial test with my PowerBook G4. With the patch installed, it would both suspend and hibernate, but WOL would be impossible. This computer uses a PCMCIA version of the BCM4318, and power is turned off to the PCMCIA card when suspended or hibernating.

Thanks for testing.

John, can you take this one? Or do we need to split the b44 part out? I added my Signed-off.

Quoted 40 lines · Michael B?sch

Quoted 36 lines · On Tue, 02 Dec 2014 16:23:49 -0600 Larry Finger <[email protected]>…

Quoted 23 lines · Michael B?sch

Quoted 20 lines · On Tue, 2 Dec 2014 23:01:29 +0300 Andrey Skvortsov…

Quoted 19 lines · Michael B?sch

Quoted 15 lines · On Mon, 1 Dec 2014 23:46:38 +0300 Andrey Skvortsov…

Wake On Lan was not working on laptop DELL Vostro 1500.
If WOL was turned on, BCM4401 was powered up in suspend mode. LEDs blinked. But the laptop could not be woken up with the Magic Packet. The reason for that was that PCIE was not enabled as a system wakeup source and
therefore the host PCI bridge was not powered up in suspend mode.
PCIE was not enabled in suspend by PM because no child devices were registered as wakeup source during suspend process.
On laptop BCM4401 is connected through the SSB bus, that is connected to the PCI-Express bus. SSB and B44 did not use standard PM wakeup functions
and did not forward wakeup settings to their parents.
To fix that B44 driver enables PM wakeup and registers new wakeup source using device_set_wakeup_enable(). Wakeup is automatically reported to the parent SSB bus via power.wakeup_path. SSB bus enables wakeup for the parent PCI bridge, if there is any child devices with enabled wakeup functionality. All other steps are
done by PM core code.

Quoted 7 lines

Quoted 4 lines

Quoted 2 lines

Thanks, this looks good.
I assume you tested this (I currently don't have a device to test this).

Sure, I've tested it. WOL from suspend is working and after resume from hibernate Ethernet is working too.

That sounds good, indeed.
I'd still prefer, if someone with b43 (wireless) would test it, too.

I did a partial test with my PowerBook G4. With the patch installed, it would both suspend and hibernate, but WOL would be impossible. This computer uses a PCMCIA version of the BCM4318, and power is turned off to the PCMCIA card when suspended or hibernating.

Quoted 4 lines

Thanks for testing.

John, can you take this one? Or do we need to split the b44 part out? I added my Signed-off.

Um, sure...3.19 is OK I presume?

Quoted 8 lines

Quoted 8 lines

Quoted 7 lines

Quoted 2 lines

That sounds good, indeed.
I'd still prefer, if someone with b43 (wireless) would test it, too.

I did a partial test with my PowerBook G4. With the patch installed, it would both suspend and hibernate, but WOL would be impossible. This computer uses a PCMCIA version of the BCM4318, and power is turned off to the PCMCIA card when suspended or hibernating.

Quoted 6 lines

Quoted 4 lines

Thanks for testing.

John, can you take this one? Or do we need to split the b44 part out? I added my Signed-off.

Um, sure...3.19 is OK I presume?

I think we could even wait for 3.20.