[PATCH]eepro100 resume failure

[PATCH]eepro100 resume failure

[Li Shaohua]

[-- Attachment #1: Type: text/plain, Size: 1095 bytes --]

Hi,
After resume from S3, eepro100 driver reported some errors and stop working in my Toshiba
laptop. Below patch fixes it.

Thanks,
Shaohua


Signed-off-by: Li Shaohua<shaohua.li@intel.com>
---

 2.6-root/drivers/net/eepro100.c |    7 ++++++-
 1 files changed, 6 insertions(+), 1 deletion(-)

diff -puN drivers/net/eepro100.c~eepro100-pm drivers/net/eepro100.c
--- 2.6/drivers/net/eepro100.c~eepro100-pm	2004-11-15 16:06:22.500880880 +0800
+++ 2.6-root/drivers/net/eepro100.c	2004-11-15 16:16:43.057541928 +0800
@@ -2327,7 +2327,7 @@ static int eepro100_suspend(struct pci_d
 	netif_device_detach(dev);
 	outl(PortPartialReset, ioaddr + SCBPort);
 	
-	/* XXX call pci_set_power_state ()? */
+	pci_set_power_state (pdev, 3);
 	return 0;
 }
 
@@ -2337,7 +2337,12 @@ static int eepro100_resume(struct pci_de
 	struct speedo_private *sp = netdev_priv(dev);
 	long ioaddr = dev->base_addr;
 
+	pci_set_power_state(pdev, 0);
 	pci_restore_state(pdev);
+	if (pdev->is_enabled)
+		pci_enable_device(pdev);
+	if (pdev->is_busmaster)
+		pci_set_master(pdev);
 
 	if (!netif_running(dev))
 		return 0;
_


[-- Attachment #2: eepro100-pm.patch --]
[-- Type: text/x-patch, Size: 905 bytes --]



---

 2.6-root/drivers/net/eepro100.c |    7 ++++++-
 1 files changed, 6 insertions(+), 1 deletion(-)

diff -puN drivers/net/eepro100.c~eepro100-pm drivers/net/eepro100.c
--- 2.6/drivers/net/eepro100.c~eepro100-pm	2004-11-15 16:06:22.500880880 +0800
+++ 2.6-root/drivers/net/eepro100.c	2004-11-15 16:16:43.057541928 +0800
@@ -2327,7 +2327,7 @@ static int eepro100_suspend(struct pci_d
 	netif_device_detach(dev);
 	outl(PortPartialReset, ioaddr + SCBPort);
 	
-	/* XXX call pci_set_power_state ()? */
+	pci_set_power_state (pdev, 3);
 	return 0;
 }
 
@@ -2337,7 +2337,12 @@ static int eepro100_resume(struct pci_de
 	struct speedo_private *sp = netdev_priv(dev);
 	long ioaddr = dev->base_addr;
 
+	pci_set_power_state(pdev, 0);
 	pci_restore_state(pdev);
+	if (pdev->is_enabled)
+		pci_enable_device(pdev);
+	if (pdev->is_busmaster)
+		pci_set_master(pdev);
 
 	if (!netif_running(dev))
 		return 0;
_

Re: [PATCH]eepro100 resume failure

[Arjan van de Ven]

>  
> @@ -2337,7 +2337,12 @@ static int eepro100_resume(struct pci_de
>  	struct speedo_private *sp = netdev_priv(dev);
>  	long ioaddr = dev->base_addr;
>  
> +	pci_set_power_state(pdev, 0);
>  	pci_restore_state(pdev);
> +	if (pdev->is_enabled)
> +		pci_enable_device(pdev);
> +	if (pdev->is_busmaster)
> +		pci_set_master(pdev);

this is wrong; the driver should KNOW the device is enabled; no reason
to check for is_enabled. Same for is_busmaster...

Re: [PATCH]eepro100 resume failure

[Li Shaohua]

[-- Attachment #1: Type: text/plain, Size: 1480 bytes --]

On Mon, 2004-11-15 at 16:43, Arjan van de Ven wrote:
> >  
> > @@ -2337,7 +2337,12 @@ static int eepro100_resume(struct pci_de
> >  	struct speedo_private *sp = netdev_priv(dev);
> >  	long ioaddr = dev->base_addr;
> >  
> > +	pci_set_power_state(pdev, 0);
> >  	pci_restore_state(pdev);
> > +	if (pdev->is_enabled)
> > +		pci_enable_device(pdev);
> > +	if (pdev->is_busmaster)
> > +		pci_set_master(pdev);
> 
> this is wrong; the driver should KNOW the device is enabled; no reason
> to check for is_enabled. Same for is_busmaster...
Oh, yes. The check isn't required. Updated one.

Thanks,
Shaohua

---

 2.6-root/drivers/net/eepro100.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletion(-)

diff -puN drivers/net/eepro100.c~eepro100-pm drivers/net/eepro100.c
--- 2.6/drivers/net/eepro100.c~eepro100-pm	2004-11-15 16:06:22.500880880 +0800
+++ 2.6-root/drivers/net/eepro100.c	2004-11-15 16:40:47.211997080 +0800
@@ -2327,7 +2327,8 @@ static int eepro100_suspend(struct pci_d
 	netif_device_detach(dev);
 	outl(PortPartialReset, ioaddr + SCBPort);
 	
-	/* XXX call pci_set_power_state ()? */
+	pci_disable_device(pdev);
+	pci_set_power_state (pdev, 3);
 	return 0;
 }
 
@@ -2337,7 +2338,10 @@ static int eepro100_resume(struct pci_de
 	struct speedo_private *sp = netdev_priv(dev);
 	long ioaddr = dev->base_addr;
 
+	pci_set_power_state(pdev, 0);
 	pci_restore_state(pdev);
+	pci_enable_device(pdev);
+	pci_set_master(pdev);
 
 	if (!netif_running(dev))
 		return 0;
_


[-- Attachment #2: eepro100-pm.patch --]
[-- Type: text/x-patch, Size: 880 bytes --]



---

 2.6-root/drivers/net/eepro100.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletion(-)

diff -puN drivers/net/eepro100.c~eepro100-pm drivers/net/eepro100.c
--- 2.6/drivers/net/eepro100.c~eepro100-pm	2004-11-15 16:06:22.500880880 +0800
+++ 2.6-root/drivers/net/eepro100.c	2004-11-15 16:40:47.211997080 +0800
@@ -2327,7 +2327,8 @@ static int eepro100_suspend(struct pci_d
 	netif_device_detach(dev);
 	outl(PortPartialReset, ioaddr + SCBPort);
 	
-	/* XXX call pci_set_power_state ()? */
+	pci_disable_device(pdev);
+	pci_set_power_state (pdev, 3);
 	return 0;
 }
 
@@ -2337,7 +2338,10 @@ static int eepro100_resume(struct pci_de
 	struct speedo_private *sp = netdev_priv(dev);
 	long ioaddr = dev->base_addr;
 
+	pci_set_power_state(pdev, 0);
 	pci_restore_state(pdev);
+	pci_enable_device(pdev);
+	pci_set_master(pdev);
 
 	if (!netif_running(dev))
 		return 0;
_

Re: [PATCH]eepro100 resume failure

[Li Shaohua]

On Mon, 2004-11-15 at 16:58, Arjan van de Ven wrote:
> > +	pci_set_power_state(pdev, 0);
> >  	pci_restore_state(pdev);
> > +	pci_enable_device(pdev);
> > +	pci_set_master(pdev);
> >  
> 
> one more nitpick; pci_enable_device() already calls
> pci_set_power_state()... isn't that double up now ?
> 
I think 'pci_restore_state' should be called after
'pci_set_power_state'.

Thanks,
Shaohua

Re: [PATCH]eepro100 resume failure

[Arjan van de Ven]

> +	pci_set_power_state(pdev, 0);
>  	pci_restore_state(pdev);
> +	pci_enable_device(pdev);
> +	pci_set_master(pdev);
>  

one more nitpick; pci_enable_device() already calls
pci_set_power_state()... isn't that double up now ?