[PATCH] r8169: do not bring device down when suspending

[PATCH] r8169: do not bring device down when suspending

[Francois Romieu]

Stopping all activity through ChipCmd and blindly acking the irqs
is neither nice nor completely needed: the transition to low-power
mode does enough work and it apparently keeps the device in a sane
state.

Patch suggested by a fix for http://bugzilla.kernel.org/show_bug.cgi?id=9512

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Tested-by: Anders Eriksson <aeriksson@fastmail.fm>
Cc: Edward Hsu <edward_hsu@realtek.com.tw>
---
 drivers/net/r8169.c |    9 ---------
 1 files changed, 0 insertions(+), 9 deletions(-)

diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c
index 3b19e0c..d354a3a 100644
--- a/drivers/net/r8169.c
+++ b/drivers/net/r8169.c
@@ -3808,7 +3808,6 @@ static int rtl8169_suspend(struct pci_dev *pdev, pm_message_t state)
 {
 	struct net_device *dev = pci_get_drvdata(pdev);
 	struct rtl8169_private *tp = netdev_priv(dev);
-	void __iomem *ioaddr = tp->mmio_addr;
 
 	if (!netif_running(dev))
 		goto out_pci_suspend;
@@ -3816,14 +3815,6 @@ static int rtl8169_suspend(struct pci_dev *pdev, pm_message_t state)
 	netif_device_detach(dev);
 	netif_stop_queue(dev);
 
-	spin_lock_irq(&tp->lock);
-
-	rtl8169_asic_down(ioaddr);
-
-	rtl8169_rx_missed(dev, ioaddr);
-
-	spin_unlock_irq(&tp->lock);
-
 out_pci_suspend:
 	pci_save_state(pdev);
 	pci_enable_wake(pdev, pci_choose_state(pdev, state),
-- 
1.6.0.6

Re: [PATCH] r8169: do not bring device down when suspending

[David Miller]

From: Francois Romieu <romieu@fr.zoreil.com>
Date: Thu, 11 Jun 2009 22:48:51 +0200

> Stopping all activity through ChipCmd and blindly acking the irqs
> is neither nice nor completely needed: the transition to low-power
> mode does enough work and it apparently keeps the device in a sane
> state.
> 
> Patch suggested by a fix for http://bugzilla.kernel.org/show_bug.cgi?id=9512
> 
> Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
> Tested-by: Anders Eriksson <aeriksson@fastmail.fm>

This patch does not apply to net-next-2.6

You're patching rtl8169_suspend(), but in net-next-2.6 it does nothing
but call a function named rtl8169_net_suspend() which actually has the
code you seem to be changing.

Please submit a fixed up patch and let me know if you want me to queue
this up to -stable.

Thanks.

Re: [PATCH] r8169: do not bring device down when suspending

[Francois Romieu]

Anders, can you try davem's net-next +

--- a/drivers/net/r8169.c
+++ b/drivers/net/r8169.c
@@ -3812,22 +3812,11 @@ static struct net_device_stats *rtl8169_get_stats(struct net_device *dev)
 
 static void rtl8169_net_suspend(struct net_device *dev)
 {
-	struct rtl8169_private *tp = netdev_priv(dev);
-	void __iomem *ioaddr = tp->mmio_addr;
-
 	if (!netif_running(dev))
 		return;
 
 	netif_device_detach(dev);
 	netif_stop_queue(dev);
-
-	spin_lock_irq(&tp->lock);
-
-	rtl8169_asic_down(ioaddr);
-
-	rtl8169_rx_missed(dev, ioaddr);
-
-	spin_unlock_irq(&tp->lock);
 }
 
 #ifdef CONFIG_PM
@@ -3877,9 +3866,17 @@ static struct dev_pm_ops rtl8169_pm_ops = {
 static void rtl_shutdown(struct pci_dev *pdev)
 {
 	struct net_device *dev = pci_get_drvdata(pdev);
+	struct rtl8169_private *tp = netdev_priv(dev);
+	void __iomem *ioaddr = tp->mmio_addr;
 
 	rtl8169_net_suspend(dev);
 
+	spin_lock_irq(&tp->lock);
+
+	rtl8169_asic_down(ioaddr);
+
+	spin_unlock_irq(&tp->lock);
+
 	if (system_state == SYSTEM_POWER_OFF) {
 		pci_wake_from_d3(pdev, true);
 		pci_set_power_state(pdev, PCI_D3hot);


If it is easier for you, you can try against 2.6.30 the patch at
http://userweb.kernel.org/~romieu/r8169/2.6.30/20090612-r8169-test.patch

Thanks.

-- 
Ueimor

Re: [PATCH] r8169: do not bring device down when suspending

[Anders Eriksson]

romieu@fr.zoreil.com said:
> If it is easier for you, you can try against 2.6.30 the patch at 
> http://userweb.kernel.org/~romieu/r8169/2.6.30/20090612-r8169-test.patch 

tested, and it does wake up on broadcasted and unicasted UDP magic packet, but
not any other kinds of packets. More specifically:

(Magic payload, UDP, IP dest link broadcast, eth dest broadcast) ->
wakes up on 'g' and 'b'

(Magic payload, UDP, IP dest unicast, eth dest unicast) ->
wakes up on 'g' and 'b'

(arp: eth broadcast) ->
does NOT wake up on 'b' as expected

Hmmm, maybe 'b' stands for IP broadcast and not eth broadcast? Let's see:
(ping broadcast: icmp,  IP dest broadcast, eth dest broadcast) ->
does NOT wake up on 'b'.

As best I can tell, the patch fixes the g case, but something's still broken 
for the other cases the device claims to support ('pumb').

-Anders

[PATCH] r8169: do not bring device down when suspending

[Francois Romieu]

Stopping all activity through ChipCmd and blindly acking the irqs
is neither nice nor completely needed: the transition to low-power
mode does enough work and it apparently keeps the device in a sane
state.

Patch suggested by a fix for http://bugzilla.kernel.org/show_bug.cgi?id=9512

The rtl_shutdown path is kept unchanged so far.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Tested-by: Anders Eriksson <aeriksson@fastmail.fm>
Cc: Edward Hsu <edward_hsu@realtek.com.tw>
---
 drivers/net/r8169.c |   19 ++++++++-----------
 1 files changed, 8 insertions(+), 11 deletions(-)

diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c
index 35196fa..4e22462 100644
--- a/drivers/net/r8169.c
+++ b/drivers/net/r8169.c
@@ -3811,22 +3811,11 @@ static struct net_device_stats *rtl8169_get_stats(struct net_device *dev)
 
 static void rtl8169_net_suspend(struct net_device *dev)
 {
-	struct rtl8169_private *tp = netdev_priv(dev);
-	void __iomem *ioaddr = tp->mmio_addr;
-
 	if (!netif_running(dev))
 		return;
 
 	netif_device_detach(dev);
 	netif_stop_queue(dev);
-
-	spin_lock_irq(&tp->lock);
-
-	rtl8169_asic_down(ioaddr);
-
-	rtl8169_rx_missed(dev, ioaddr);
-
-	spin_unlock_irq(&tp->lock);
 }
 
 #ifdef CONFIG_PM
@@ -3876,9 +3865,17 @@ static struct dev_pm_ops rtl8169_pm_ops = {
 static void rtl_shutdown(struct pci_dev *pdev)
 {
 	struct net_device *dev = pci_get_drvdata(pdev);
+	struct rtl8169_private *tp = netdev_priv(dev);
+	void __iomem *ioaddr = tp->mmio_addr;
 
 	rtl8169_net_suspend(dev);
 
+	spin_lock_irq(&tp->lock);
+
+	rtl8169_asic_down(ioaddr);
+
+	spin_unlock_irq(&tp->lock);
+
 	if (system_state == SYSTEM_POWER_OFF) {
 		pci_wake_from_d3(pdev, true);
 		pci_set_power_state(pdev, PCI_D3hot);
-- 
1.6.2.2

Re: [PATCH] r8169: do not bring device down when suspending

[David Miller]

From: Francois Romieu <romieu@fr.zoreil.com>
Date: Wed, 17 Jun 2009 23:41:45 +0200

> Stopping all activity through ChipCmd and blindly acking the irqs
> is neither nice nor completely needed: the transition to low-power
> mode does enough work and it apparently keeps the device in a sane
> state.
> 
> Patch suggested by a fix for http://bugzilla.kernel.org/show_bug.cgi?id=9512
> 
> The rtl_shutdown path is kept unchanged so far.
> 
> Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
> Tested-by: Anders Eriksson <aeriksson@fastmail.fm>

Applied.