Regression in 2.6.27-rc7: Wake On LAN with sky2 broken
Subsystems (2)
Hi,
Wake On LAN from suspend to RAM doesn't work on my Mac mini Core Duo with 2.6.27-rc7. I tested the same configuration with 2.6.26.3, and it worked. With 2.6.27-rc7, the computer stays in suspended state. I can still wake it up using the keyboard or power switch.
In my suspend script, I use the command "ethtool -s eth0 wol g" to enable WOL even if it was disabled before.
Any hints what commit I should try to revert, before I start bisecting?
Regards,
Tino
Quoted 11 lines · Tino Keitel
Hi,
Wake On LAN from suspend to RAM doesn't work on my Mac mini Core Duo with 2.6.27-rc7. I tested the same configuration with 2.6.26.3, and it worked. With 2.6.27-rc7, the computer stays in suspended state. I can still wake it up using the keyboard or power switch.
In my suspend script, I use the command "ethtool -s eth0 wol g" to enable WOL even if it was disabled before.
Any hints what commit I should try to revert, before I start bisecting?
Please try the appended patch.
Thanks,
Rafael
---
From: Rafael J. Wysocki <[email protected]>
sky2: Fix wake-on-LAN regression
* Since dev->power.should_wakeup bit is used by the PCI core to decide whether the device should wake up the system from sleep states, set/unset this bit whenever WOL is enabled/disabled using sky2_set_wol().
* Use pci_prepare_to_sleep() and pci_back_from_sleep() in ->sky2_suspend() and ->sky2_resume(), respectively, instead of the older less suitable interfaces.
* Prevent pci_enable_wake() from being called twice unnecessarily in sky2_shutdown().
---
drivers/net/sky2.c | 34 ++++++++++------------------------ 1 file changed, 10 insertions(+), 24 deletions(-)
drivers/net/sky2.c renamed from drivers/net/sky2.c~skty2-adapt-to-the-reworked-pci-pm +10 -24
@@ -3034,7 +3034,8 @@ static int sky2_set_wol(struct net_devicstruct sky2_port *sky2 =netdev_priv (dev);struct sky2_hw *hw = sky2->hw; -if (wol->wolopts & ~sky2_wol_supported (sky2->hw)) +if ((wol->wolopts & ~sky2_wol_supported (sky2->hw)) + || !device_can_wakeup (&hw->pdev->dev))return -EOPNOTSUPP ; sky2->wol = wol->wolopts;
@@ -3045,6 +3046,8 @@ static int sky2_set_wol(struct net_devicsky2_write32 (hw,B0_CTST , sky2->wol ?Y2_HW_WOL_ON :Y2_HW_WOL_OFF ); +device_set_wakeup_enable (&hw->pdev->dev, sky2->wol); +if (!netif_running (dev))sky2_wol_init (sky2);return 0 ;
@@ -4179,18 +4182,6 @@ static int __devinit sky2_test_msi(strucreturn err; } -static int __devinit pci_wake_enabled (struct pci_dev *dev) -{ -int pm =pci_find_capability (dev,PCI_CAP_ID_PM ); -u16 value; - -if (!pm) -return 0 ; -if (pci_read_config_word (dev, pm +PCI_PM_CTRL , &value)) -return 0 ; -return value &PCI_PM_CTRL_PME_ENABLE ; -} -/* This driver supports yukon2 chipset only */ static const char *sky2_name (u8 chipid,char *buf,int sz) {
@@ -4314,7 +4305,7 @@ static int __devinit sky2_probe(struct p } } - wol_default =pci_wake_enabled (pdev) ?WAKE_MAGIC :0 ; + wol_default =device_may_wakeup (&pdev->dev) ?WAKE_MAGIC :0 ; err = -ENOMEM ; hw =kzalloc (sizeof (*hw),GFP_KERNEL );
@@ -4479,7 +4470,7 @@ static void __devexit sky2_remove(structstatic int sky2_suspend (struct pci_dev *pdev,pm_message_t state) {struct sky2_hw *hw =pci_get_drvdata (pdev); -int i, wol =0 ; +int i;if (!hw)return 0 ;
@@ -4497,8 +4488,6 @@ static int sky2_suspend(struct pci_dev *if (sky2->wol)sky2_wol_init (sky2); - - wol |= sky2->wol; }sky2_write32 (hw,B0_IMSK ,0 );
@@ -4506,8 +4495,7 @@ static int sky2_suspend(struct pci_dev *sky2_power_aux (hw);pci_save_state (pdev); -pci_enable_wake (pdev,pci_choose_state (pdev, state), wol); -pci_set_power_state (pdev,pci_choose_state (pdev, state)); +pci_prepare_to_sleep (pdev);return 0 ; }
@@ -4520,7 +4508,7 @@ static int sky2_resume(struct pci_dev *pif (!hw)return 0 ; - err =pci_set_power_state (pdev,PCI_D0 ); + err =pci_back_from_sleep (pdev);if (err) goto out;
@@ -4528,8 +4516,6 @@ static int sky2_resume(struct pci_dev *pif (err) goto out; -pci_enable_wake (pdev,PCI_D0 ,0 ); -/* Re-enable all clocks */ if (hw->chip_id == CHIP_ID_YUKON_EX || hw->chip_id == CHIP_ID_YUKON_EC_U ||
@@ -4588,8 +4574,8 @@ static void sky2_shutdown(struct pci_devif (wol)sky2_power_aux (hw); -pci_enable_wake (pdev, PCI_D3hot, wol); -pci_enable_wake (pdev, PCI_D3cold, wol); +if (pci_enable_wake (pdev, PCI_D3cold, wol)) +pci_enable_wake (pdev, PCI_D3hot, wol);pci_disable_device (pdev);pci_set_power_state (pdev, PCI_D3hot);
_
Quoted 8 lines · Rafael J. Wysocki
Quoted 7 lines · Tino Keitel
Hi,
Wake On LAN from suspend to RAM doesn't work on my Mac mini Core Duo with 2.6.27-rc7. I tested the same configuration with 2.6.26.3, and it worked. With 2.6.27-rc7, the computer stays in suspended state. I can still wake it up using the keyboard or power switch.
Quoted 6 lines
Quoted 4 lines
In my suspend script, I use the command "ethtool -s eth0 wol g" to enable WOL even if it was disabled before.
Any hints what commit I should try to revert, before I start bisecting?
Please try the appended patch.
With this patch, the computer hangs during suspend.
Regards,
Tino
Quoted 9 lines · Tino Keitel
Quoted 8 lines · Rafael J. Wysocki
Quoted 7 lines · Tino Keitel
Hi,
Wake On LAN from suspend to RAM doesn't work on my Mac mini Core Duo with 2.6.27-rc7. I tested the same configuration with 2.6.26.3, and it worked. With 2.6.27-rc7, the computer stays in suspended state. I can still wake it up using the keyboard or power switch.
Quoted 8 lines
Quoted 6 lines
Quoted 4 lines
In my suspend script, I use the command "ethtool -s eth0 wol g" to enable WOL even if it was disabled before.
Any hints what commit I should try to revert, before I start bisecting?
Please try the appended patch.
With this patch, the computer hangs during suspend.
This is not good. It probably means that it actually tries to set up the NIC for waking up the system, but something broke _that_ for you.
FWIW, I tried it on my test machine with sky2 and it worked.
Can you try:
# ethtool -s eth0 wol g # echo devices > /sys/power/pm_test # echo mem > /sys/power/state (this should wait for 5 seconds and return to the command line)
and see if the machine hangs during this operation?
Please also attach the boot log from your system.
Thanks,
Rafael
On Wed, Sep 24, 2008 at 23:10:22 +0200, Rafael J. Wysocki wrote:
[...]
Quoted 8 lines
Can you try:
# ethtool -s eth0 wol g # echo devices > /sys/power/pm_test # echo mem > /sys/power/state (this should wait for 5 seconds and return to the command line)and see if the machine hangs during this operation?
It didn't hang. The dmesg output from the suspend attempt is here:
Quoted 2 lines
Please also attach the boot log from your system.
dmesg output after boot:
Regards,
Tino
Hi,
I just found a method how I can enable WOL again:
echo enabled > /sys/devices/pci0000:00/0000:00:1c.0/0000:01:00.0/power/wakeup
This device is the NIC device, of cause.
I already had another case (EHCI) where I needed to do this to be able to wake the computer up with the USB keyboard. It seems like 2.6.27 behaves different here compared to 2.6.26, and things that worked with 2.6.26 need manual fixing to work again. Or did I make something wrong?
Regards,
Tino
Quoted 8 lines · Tino Keitel
Hi,
I just found a method how I can enable WOL again:
echo enabled > /sys/devices/pci0000:00/0000:00:1c.0/0000:01:00.0/power/wakeup
This device is the NIC device, of cause.
I guess the box didn't hang during suspend with this setting?
If that's correct, please test the patch below.
Quoted 4 lines
I already had another case (EHCI) where I needed to do this to be able to wake the computer up with the USB keyboard. It seems like 2.6.27 behaves different here compared to 2.6.26, and things that worked with 2.6.26 need manual fixing to work again. Or did I make something wrong?
No, you didn't.
The behavior was changed by the PCI wake-up patches that fixed a couple of problems and I'd like to fix all of the cases when the new behavior causes issues like this to happen.
I'll send you a patch for the EHCI case later today.
Thanks,
Rafael
---
sky2: Fix WOL regression
Since dev->power.should_wakeup bit is used by the PCI core to decide whether the device should wake up the system from sleep states, set/unset this bit whenever WOL is enabled/disabled using sky2_set_wol().
---
drivers/net/sky2.c | 34 ++++++++++------------------------ 1 file changed, 10 insertions(+), 24 deletions(-)
drivers/net/sky2.c renamed from drivers/net/sky2.c~skty2-adapt-to-the-reworked-pci-pm +5 -2
@@ -3034,7 +3034,8 @@ static int sky2_set_wol(struct net_devicstruct sky2_port *sky2 =netdev_priv (dev);struct sky2_hw *hw = sky2->hw; -if (wol->wolopts & ~sky2_wol_supported (sky2->hw)) +if ((wol->wolopts & ~sky2_wol_supported (sky2->hw)) + || !device_can_wakeup (&hw->pdev->dev))return -EOPNOTSUPP ; sky2->wol = wol->wolopts;
@@ -3045,6 +3046,8 @@ static int sky2_set_wol(struct net_devicsky2_write32 (hw,B0_CTST , sky2->wol ?Y2_HW_WOL_ON :Y2_HW_WOL_OFF ); +device_set_wakeup_enable (&hw->pdev->dev, sky2->wol); +if (!netif_running (dev))sky2_wol_init (sky2);return 0 ;
@@ -4314,7 +4305,7 @@ static int __devinit sky2_probe(struct p } } - wol_default =pci_wake_enabled (pdev) ?WAKE_MAGIC :0 ; + wol_default =device_may_wakeup (&pdev->dev) ?WAKE_MAGIC :0 ; err = -ENOMEM ; hw =kzalloc (sizeof (*hw),GFP_KERNEL );
Quoted 5 lines · Rafael J. Wysocki
Quoted 4 lines · Tino Keitel
Hi,
I just found a method how I can enable WOL again:
Quoted 8 lines
Quoted 4 lines
echo enabled > /sys/devices/pci0000:00/0000:00:1c.0/0000:01:00.0/power/wakeup
This device is the NIC device, of cause.
I guess the box didn't hang during suspend with this setting?
If that's correct, please test the patch below.
With the attached patch, WOL works as intended the hang also didn't happen.
Thanks and regards,
Tino