[patch 17/18] 3c59x: fix duplex configuration

[patch 17/18] 3c59x: fix duplex configuration

[akpm]

From: Steffen Klassert <klassert@mathematik.tu-chemnitz.de>

A special sequence of ifconfig up/down and plug/unplug the cable can break
the duplex configuration of the driver.

Setting
vp->mii.full_duplex = vp->full_duplex
in vortex_up should fix this.

Addresses Bug 8575 3c59x duplex configuration broken
http://bugzilla.kernel.org/show_bug.cgi?id=8575

Cc: Martin Buck <mb-tmp-ohtmvyyn.xreary.bet@gromit.dyndns.org>
Signed-off-by: Steffen Klassert <klassert@mathematik.tu-chemnitz.de>
Cc: Natalie Protasevich <protasnb@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 drivers/net/3c59x.c |    1 +
 1 files changed, 1 insertion(+)

diff -puN drivers/net/3c59x.c~3c59x-fix-duplex-configuration drivers/net/3c59x.c
--- a/drivers/net/3c59x.c~3c59x-fix-duplex-configuration
+++ a/drivers/net/3c59x.c
@@ -1559,6 +1559,7 @@ vortex_up(struct net_device *dev)
 		mii_reg1 = mdio_read(dev, vp->phys[0], MII_BMSR);
 		mii_reg5 = mdio_read(dev, vp->phys[0], MII_LPA);
 		vp->partner_flow_ctrl = ((mii_reg5 & 0x0400) != 0);
+		vp->mii.full_duplex = vp->full_duplex;
 
 		vortex_check_media(dev, 1);
 	}
_

Re: [patch 17/18] 3c59x: fix duplex configuration

[Jeff Garzik]

akpm@linux-foundation.org wrote:
> From: Steffen Klassert <klassert@mathematik.tu-chemnitz.de>
> 
> A special sequence of ifconfig up/down and plug/unplug the cable can break
> the duplex configuration of the driver.
> 
> Setting
> vp->mii.full_duplex = vp->full_duplex
> in vortex_up should fix this.
> 
> Addresses Bug 8575 3c59x duplex configuration broken
> http://bugzilla.kernel.org/show_bug.cgi?id=8575
> 
> Cc: Martin Buck <mb-tmp-ohtmvyyn.xreary.bet@gromit.dyndns.org>
> Signed-off-by: Steffen Klassert <klassert@mathematik.tu-chemnitz.de>
> Cc: Natalie Protasevich <protasnb@gmail.com>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> ---
> 
>  drivers/net/3c59x.c |    1 +
>  1 files changed, 1 insertion(+)

applied

Re: [patch 17/18] 3c59x: fix duplex configuration

[Martin Buck]

On Fri, Aug 10, 2007 at 02:05:26PM -0700, akpm@linux-foundation.org wrote:
[...]
> diff -puN drivers/net/3c59x.c~3c59x-fix-duplex-configuration drivers/net/3c59x.c
> --- a/drivers/net/3c59x.c~3c59x-fix-duplex-configuration
> +++ a/drivers/net/3c59x.c
> @@ -1559,6 +1559,7 @@ vortex_up(struct net_device *dev)
>  		mii_reg1 = mdio_read(dev, vp->phys[0], MII_BMSR);
>  		mii_reg5 = mdio_read(dev, vp->phys[0], MII_LPA);
>  		vp->partner_flow_ctrl = ((mii_reg5 & 0x0400) != 0);
> +		vp->mii.full_duplex = vp->full_duplex;
>  
>  		vortex_check_media(dev, 1);
>  	}
> _


Sorry for the late reply. I finally managed to get my notebook fixed so
that I could actually test this patch. I can confirm that it fixes my
duplex configuration problem. The steps described in
http://bugzilla.kernel.org/show_bug.cgi?id=8575
now result in a Ethernet chip properly configured for full duplex. Thanks
for the fix!

The only remaining issue I have the 3c59x driver is the time required until
it detects link loss when unplugging the Ethernet cable. At the moment,
this needs up to 60 seconds which makes this feature pretty useless. Other
drivers need 2-5 seconds for this which is roughly what I would have
expected. I've been using the patch below sucessfully for a few weeks now
which brings down this time to 5 seconds. Would be nice if somebody could
apply it.

Thanks,
Martin


--- drivers/net/3c59x.c.orig	2007-08-20 17:01:06.000000000 +0200
+++ drivers/net/3c59x.c	2007-08-20 17:02:38.000000000 +0200
@@ -1726,7 +1726,7 @@
 	struct net_device *dev = (struct net_device *)data;
 	struct vortex_private *vp = netdev_priv(dev);
 	void __iomem *ioaddr = vp->ioaddr;
-	int next_tick = 60*HZ;
+	int next_tick = 5*HZ;
 	int ok = 0;
 	int media_status, old_window;
 
@@ -1771,9 +1771,6 @@
 		ok = 1;
 	}
 
-	if (!netif_carrier_ok(dev))
-		next_tick = 5*HZ;
-
 	if (vp->medialock)
 		goto leave_media_alone;