Patchwork netdev/fec: fix ifconfig eth0 down hang issue (v3)

login
register
mail settings
Submitter Bryan Wu
Date April 22, 2010, 6:16 a.m.
Message ID <1271917018-5756-1-git-send-email-bryan.wu@canonical.com>
Download mbox | patch
Permalink /patch/50691/
State Superseded
Delegated to: Andy Whitcroft
Headers show

Comments

Bryan Wu - April 22, 2010, 6:16 a.m.
BugLink: http://bugs.launchpad.net/bugs/559065

v3:
if fec_enet_mii_probe() fails, we should exit the open and resume functions
properly and give up further operations without a proper PHY connected.

v2:
Check fec_enet_mii_probe() return value directly, instead of fep->phy_dev

v1:
In ethernet connection open/close function, we need to use phy_connect
and phy_disconnect operation before we start/stop phy. Otherwise it will
cause system hang. Also suspend/resume needs the same fix.

Signed-off-by: Bryan Wu <bryan.wu@canonical.com>
---
 drivers/net/fec.c |   21 ++++++++++++++++++++-
 1 files changed, 20 insertions(+), 1 deletions(-)
Tim Gardner - April 22, 2010, 12:57 p.m.
On 04/22/2010 12:16 AM, Bryan Wu wrote:
> BugLink: http://bugs.launchpad.net/bugs/559065
>
> v3:
> if fec_enet_mii_probe() fails, we should exit the open and resume functions
> properly and give up further operations without a proper PHY connected.
>
> v2:
> Check fec_enet_mii_probe() return value directly, instead of fep->phy_dev
>
> v1:
> In ethernet connection open/close function, we need to use phy_connect
> and phy_disconnect operation before we start/stop phy. Otherwise it will
> cause system hang. Also suspend/resume needs the same fix.
>
> Signed-off-by: Bryan Wu<bryan.wu@canonical.com>
> ---
>   drivers/net/fec.c |   21 ++++++++++++++++++++-
>   1 files changed, 20 insertions(+), 1 deletions(-)
>
> diff --git a/drivers/net/fec.c b/drivers/net/fec.c
> index 2280373..b14e4d5 100644
> --- a/drivers/net/fec.c
> +++ b/drivers/net/fec.c
> @@ -929,7 +929,13 @@ fec_enet_open(struct net_device *dev)
>   		return ret;
>
>   	/* schedule a link state check */
> +	ret = fec_enet_mii_probe(dev);
> +	if (!ret) {
> +		fec_enet_free_buffers(dev);
> +		return ret;
> +	}
>   	phy_start(fep->phy_dev);
> +
>   	netif_start_queue(dev);
>   	fep->opened = 1;
>   	return 0;
> @@ -942,10 +948,13 @@ fec_enet_close(struct net_device *dev)
>
>   	/* Don't know what to do yet. */
>   	fep->opened = 0;
> -	phy_stop(fep->phy_dev);
>   	netif_stop_queue(dev);
> +
>   	fec_stop(dev);
>
> +	if (fep->phy_dev)
> +		phy_disconnect(fep->phy_dev);
> +
>           fec_enet_free_buffers(dev);
>   	clk_disable(fep->clk);
>
> @@ -1396,6 +1405,8 @@ fec_suspend(struct platform_device *dev, pm_message_t state)
>   		fep = netdev_priv(ndev);
>   		if (netif_running(ndev)) {
>   			netif_device_detach(ndev);
> +			if (fep->phy_dev)
> +				phy_disconnect(fep->phy_dev);
>   			fec_stop(ndev);
>   			clk_disable(fep->clk);
>   		}
> @@ -1409,12 +1420,20 @@ fec_resume(struct platform_device *dev)
>   {
>   	struct net_device *ndev = platform_get_drvdata(dev);
>   	struct fec_enet_private *fep;
> +	int ret;
>
>   	if (ndev) {
>   		fep = netdev_priv(ndev);
>   		if (netif_running(ndev)) {
>   			clk_enable(fep->clk);
>   			fec_restart(ndev, 1);
> +			ret = fec_enet_mii_probe(ndev);
> +			if (!ret) {
> +				fec_stop(ndev);
> +				clk_disable(fep->clk);
> +				return ret;
> +			}
> +			phy_start(fep->phy_dev);
>   			netif_device_attach(ndev);
>   		}
>   	}

One final comment. fec_enet_mii_probe() should set fep->phy_dev=NULL 
upon entry so that if it errors out fep->phy_dev is not left with a 
bogus value (which could happen in fec_resume()).

How is the fsl-imx51 physically constructed wrt to the ethernet phy? Is 
it possible for the phy to disappear during a suspend/resume cycle?

rtg
Bryan Wu - April 22, 2010, 2:24 p.m.
On 04/22/2010 08:57 PM, Tim Gardner wrote:
> On 04/22/2010 12:16 AM, Bryan Wu wrote:
>> BugLink: http://bugs.launchpad.net/bugs/559065
>>
>> v3:
>> if fec_enet_mii_probe() fails, we should exit the open and resume
>> functions
>> properly and give up further operations without a proper PHY connected.
>>
>> v2:
>> Check fec_enet_mii_probe() return value directly, instead of fep->phy_dev
>>
>> v1:
>> In ethernet connection open/close function, we need to use phy_connect
>> and phy_disconnect operation before we start/stop phy. Otherwise it will
>> cause system hang. Also suspend/resume needs the same fix.
>>
>> Signed-off-by: Bryan Wu<bryan.wu@canonical.com>
>> ---
>> drivers/net/fec.c | 21 ++++++++++++++++++++-
>> 1 files changed, 20 insertions(+), 1 deletions(-)
>>
>> diff --git a/drivers/net/fec.c b/drivers/net/fec.c
>> index 2280373..b14e4d5 100644
>> --- a/drivers/net/fec.c
>> +++ b/drivers/net/fec.c
>> @@ -929,7 +929,13 @@ fec_enet_open(struct net_device *dev)
>> return ret;
>>
>> /* schedule a link state check */
>> + ret = fec_enet_mii_probe(dev);
>> + if (!ret) {
>> + fec_enet_free_buffers(dev);
>> + return ret;
>> + }
>> phy_start(fep->phy_dev);
>> +
>> netif_start_queue(dev);
>> fep->opened = 1;
>> return 0;
>> @@ -942,10 +948,13 @@ fec_enet_close(struct net_device *dev)
>>
>> /* Don't know what to do yet. */
>> fep->opened = 0;
>> - phy_stop(fep->phy_dev);
>> netif_stop_queue(dev);
>> +
>> fec_stop(dev);
>>
>> + if (fep->phy_dev)
>> + phy_disconnect(fep->phy_dev);
>> +
>> fec_enet_free_buffers(dev);
>> clk_disable(fep->clk);
>>
>> @@ -1396,6 +1405,8 @@ fec_suspend(struct platform_device *dev,
>> pm_message_t state)
>> fep = netdev_priv(ndev);
>> if (netif_running(ndev)) {
>> netif_device_detach(ndev);
>> + if (fep->phy_dev)
>> + phy_disconnect(fep->phy_dev);
>> fec_stop(ndev);
>> clk_disable(fep->clk);
>> }
>> @@ -1409,12 +1420,20 @@ fec_resume(struct platform_device *dev)
>> {
>> struct net_device *ndev = platform_get_drvdata(dev);
>> struct fec_enet_private *fep;
>> + int ret;
>>
>> if (ndev) {
>> fep = netdev_priv(ndev);
>> if (netif_running(ndev)) {
>> clk_enable(fep->clk);
>> fec_restart(ndev, 1);
>> + ret = fec_enet_mii_probe(ndev);
>> + if (!ret) {
>> + fec_stop(ndev);
>> + clk_disable(fep->clk);
>> + return ret;
>> + }
>> + phy_start(fep->phy_dev);
>> netif_device_attach(ndev);
>> }
>> }
>
> One final comment. fec_enet_mii_probe() should set fep->phy_dev=NULL
> upon entry so that if it errors out fep->phy_dev is not left with a
> bogus value (which could happen in fec_resume()).
>

Yeah, good point. I will provide another patch to add this.

> How is the fsl-imx51 physically constructed wrt to the ethernet phy? Is
> it possible for the phy to disappear during a suspend/resume cycle?
>

Actually, FEC is an on-chip controller of imx51 and PHY is an external chip 
which connected to imx51 via MDIO and MII bus. Since suspend/resume operation is 
very similar to close/open thing here, I just wanna keep them similar.

Thanks,

Patch

diff --git a/drivers/net/fec.c b/drivers/net/fec.c
index 2280373..b14e4d5 100644
--- a/drivers/net/fec.c
+++ b/drivers/net/fec.c
@@ -929,7 +929,13 @@  fec_enet_open(struct net_device *dev)
 		return ret;
 
 	/* schedule a link state check */
+	ret = fec_enet_mii_probe(dev);
+	if (!ret) {
+		fec_enet_free_buffers(dev);
+		return ret;
+	}
 	phy_start(fep->phy_dev);
+
 	netif_start_queue(dev);
 	fep->opened = 1;
 	return 0;
@@ -942,10 +948,13 @@  fec_enet_close(struct net_device *dev)
 
 	/* Don't know what to do yet. */
 	fep->opened = 0;
-	phy_stop(fep->phy_dev);
 	netif_stop_queue(dev);
+
 	fec_stop(dev);
 
+	if (fep->phy_dev)
+		phy_disconnect(fep->phy_dev);
+
         fec_enet_free_buffers(dev);
 	clk_disable(fep->clk);
 
@@ -1396,6 +1405,8 @@  fec_suspend(struct platform_device *dev, pm_message_t state)
 		fep = netdev_priv(ndev);
 		if (netif_running(ndev)) {
 			netif_device_detach(ndev);
+			if (fep->phy_dev)
+				phy_disconnect(fep->phy_dev);
 			fec_stop(ndev);
 			clk_disable(fep->clk);
 		}
@@ -1409,12 +1420,20 @@  fec_resume(struct platform_device *dev)
 {
 	struct net_device *ndev = platform_get_drvdata(dev);
 	struct fec_enet_private *fep;
+	int ret;
 
 	if (ndev) {
 		fep = netdev_priv(ndev);
 		if (netif_running(ndev)) {
 			clk_enable(fep->clk);
 			fec_restart(ndev, 1);
+			ret = fec_enet_mii_probe(ndev);
+			if (!ret) {
+				fec_stop(ndev);
+				clk_disable(fep->clk);
+				return ret;
+			}
+			phy_start(fep->phy_dev);
 			netif_device_attach(ndev);
 		}
 	}