[PATCH] ibm_emac: Correctly detect old link speed

This patch fixes a bug where the link speed change was not
detected correctly. This occured on a 440SPe (EMAC4) system
where the old link speed was 100Mbps and the new link speed
is 1000Mbps.

Signed-off-by: Stefan Roese <sr@denx.de>

---
commit 1e2a6085bbb6bd384e0812b6e9c62d18d14e1c0f
tree f441a914c9d66fb00fcde1e4bd7e02effba20dc7
parent 7b104bcb8e460e45a1aebe3da9b86aacdb4cab12
author Stefan Roese [off-list ref] Wed, 16 May 2007 12:54:40 +0200
committer Stefan Roese [off-list ref] Wed, 16 May 2007 12:54:40 +0200

 drivers/net/ibm_emac/ibm_emac_core.c |   18 ++++++++++++------
 1 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ibm_emac/ibm_emac_core.c b/drivers/net/ibm_emac/ibm_emac_core.c
index 50035eb..ab38727 100644
--- a/drivers/net/ibm_emac/ibm_emac_core.c
+++ b/drivers/net/ibm_emac/ibm_emac_core.c

@@ -926,12 +926,18 @@ static int emac_link_differs(struct ocp_enet_private *dev)
 	int duplex = r & EMAC_MR1_FDE ? DUPLEX_FULL : DUPLEX_HALF;
 	int speed, pause, asym_pause;
 
-	if (r & (EMAC_MR1_MF_1000 | EMAC_MR1_MF_1000GPCS))
-		speed = SPEED_1000;
-	else if (r & EMAC_MR1_MF_100)
-		speed = SPEED_100;
-	else
-		speed = SPEED_10;
+	switch (r & EMAC_MR1_MF_MASK) {
+	case EMAC_MR1_MF_1000:
+	case EMAC_MR1_MF_1000GPCS:
+		speed = 1000;
+		break;
+	case EMAC_MR1_MF_10:
+		speed = 10;
+		break;
+	case EMAC_MR1_MF_100:
+	default:
+		speed = 100;
+	}
 
 	switch (r & (EMAC_MR1_EIFC | EMAC_MR1_APP)) {
 	case (EMAC_MR1_EIFC | EMAC_MR1_APP):

On Wed, May 16, 2007 at 01:00:08PM +0200, Stefan Roese wrote:

This patch fixes a bug where the link speed change was not
detected correctly. This occured on a 440SPe (EMAC4) system
where the old link speed was 100Mbps and the new link speed
is 1000Mbps.

Good catch, Stefan. Unfortunately, I have to NACK your patch - you 
broke non EMAC4 builds.

Correct fix is just to remove EMAC_MR1_MF_1000GPCS from the first 
if condition.

I'll send correct fix shortly along with other queued patches.

-- 
Eugene

On Wednesday 16 May 2007, Eugene Surovegin wrote:

On Wed, May 16, 2007 at 01:00:08PM +0200, Stefan Roese wrote:

quoted

This patch fixes a bug where the link speed change was not
detected correctly. This occured on a 440SPe (EMAC4) system
where the old link speed was 100Mbps and the new link speed
is 1000Mbps.

Good catch, Stefan. Unfortunately, I have to NACK your patch - you
broke non EMAC4 builds.

Yes, you're right of course.

Correct fix is just to remove EMAC_MR1_MF_1000GPCS from the first
if condition.

Yep.

I'll send correct fix shortly along with other queued patches.

Thanks.

Best regards,
Stefan

On Wed, May 16, 2007 at 08:44:47PM +0200, Stefan Roese wrote:

On Wednesday 16 May 2007, Eugene Surovegin wrote:

quoted

On Wed, May 16, 2007 at 01:00:08PM +0200, Stefan Roese wrote:

quoted

This patch fixes a bug where the link speed change was not
detected correctly. This occured on a 440SPe (EMAC4) system
where the old link speed was 100Mbps and the new link speed
is 1000Mbps.

Good catch, Stefan. Unfortunately, I have to NACK your patch - you
broke non EMAC4 builds.

Yes, you're right of course.

quoted

Correct fix is just to remove EMAC_MR1_MF_1000GPCS from the first
if condition.

Yep.

quoted

I'll send correct fix shortly along with other queued patches.

Thanks.

I've merged essentially the same fix into the device tree aware
"new_emac" driver.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson