[PATCH] via-rhine.c, wol-bugfix, Kernel 2.6.11.5

[PATCH] via-rhine.c, wol-bugfix, Kernel 2.6.11.5

[Thomas Skora]

[-- Attachment #1: Type: text/plain, Size: 246 bytes --]

Hello!

The via-rhine driver in the actual kernel release 2.6.11.5 resets
wake-on-lan-settings of the chip. This leads to the fact, that wol is
disabled after the first reboot. I've attached a little patch, that
fixes the problem.

Thomas Skora


[-- Attachment #2: patch-via-wol --]
[-- Type: application/octet-stream, Size: 509 bytes --]

--- linux-2.6.11.5/drivers/net/via-rhine.c.orig	2005-03-24 00:54:16.000000000 +0100
+++ linux-2.6.11.5/drivers/net/via-rhine.c	2005-03-24 00:54:16.000000000 +0100
@@ -542,7 +542,7 @@ static void rhine_power_init(struct net_
 
 	if (rp->quirks & rqWOL) {
 		/* Make sure chip is in power state D0 */
-		iowrite8(ioread8(ioaddr + StickyHW) & 0xFE, ioaddr + StickyHW);
+		iowrite8(ioread8(ioaddr + StickyHW) & 0xFC, ioaddr + StickyHW);
 
 		/* Disable "force PME-enable" */
 		iowrite8(0x80, ioaddr + WOLcgClr);

Re: [PATCH] via-rhine.c, wol-bugfix, Kernel 2.6.11.5

[Roger Luethi]

On Thu, 24 Mar 2005 01:37:55 +0100, Thomas Skora wrote:
> The via-rhine driver in the actual kernel release 2.6.11.5 resets
> wake-on-lan-settings of the chip. This leads to the fact, that wol is
> disabled after the first reboot. I've attached a little patch, that
> fixes the problem.

This patch won't apply to the 2.6.11.5 I have here -- the driver is
clearing ~0xFC already. The description makes me suspect that the patch
is meant to go the other way, and that would be wrong. Use ethtool if
you want to enable WOL, default in Linux net drivers is off.

Please don't send patches as application/octet-stream, that's
annoying. Send either text/plain or inline.

Roger

Re: [PATCH] via-rhine.c, wol-bugfix, Kernel 2.6.11.5

[Thomas Skora]

Roger Luethi <rl@hellgate.ch> writes:

> This patch won't apply to the 2.6.11.5 I have here -- the driver is
> clearing ~0xFC already. The description makes me suspect that the patch
> is meant to go the other way, and that would be wrong.

You are right. I don't know what happened, but in my sources (from
kernel.org without any patches) this was cleared to 0xFE.

> Use ethtool if you want to enable WOL, default in Linux net drivers is
> off.

The problem was, that this doesn't worked.

> Please don't send patches as application/octet-stream, that's
> annoying. Send either text/plain or inline.

Ok, thank you for the hint.

Thomas