2004-06-01 21:44:02

by Pavel Machek

Subject: [trivial] no mili-bits-per-second for via-rhine

Hi!

via-rhine claims to support 100 mili-bits-per-second mode and one
place and 100 mega-bit-second mode ("100 mega-bit-seconds of storage
for only $1?"). This cleans it up.

Pavel

--- tmp/linux/drivers/net/via-rhine.c 2004-05-20 23:08:19.000000000 +0200
+++ linux/drivers/net/via-rhine.c 2004-05-20 23:11:26.000000000 +0200
@@ -863,12 +863,12 @@
if (np->default_port & 0x330) {
/* FIXME: shouldn't someone check this variable? */
/* np->medialock = 1; */
- printk(KERN_INFO " Forcing %dMbs %s-duplex operation.\n",
+ printk(KERN_INFO " Forcing %dMbps %s-duplex operation.\n",
(option & 0x300 ? 100 : 10),
(option & 0x220 ? "full" : "half"));
if (np->mii_cnt)
mdio_write(dev, np->phys[0], MII_BMCR,
- ((option & 0x300) ? 0x2000 : 0) | /* 100mbps? */
+ ((option & 0x300) ? 0x2000 : 0) | /* 100Mbps? */
((option & 0x220) ? 0x0100 : 0)); /* Full duplex? */
}
}

--
934a471f20d6580d5aad759bf0d97ddc

2004-06-02 08:41:16

by Roger Luethi

Subject: Re: [trivial] no mili-bits-per-second for via-rhine

FYI, I have a patch for 2.6 in my out queue which deletes the
option/full_duplex code (including the part below) entirely. Nobody
complained when it was unusable for the longest time (kind of fixed
recently), and the code is still a mess. People using 2.6 should use
ethtool _if_ they need to force media, because that works.

So as via-rhine maintainer, I recommend applying to 2.4 and leaving
2.6 as is.

FWIW cleaning up all the printed messages in via-rhine is on my todo
list.

Roger

On Tue, 01 Jun 2004 23:43:45 +0200, Pavel Machek wrote:
> Hi!
>
> via-rhine claims to support 100 mili-bits-per-second mode and one
> place and 100 mega-bit-second mode ("100 mega-bit-seconds of storage
> for only $1?"). This cleans it up.
>
> Pavel
>
> --- tmp/linux/drivers/net/via-rhine.c 2004-05-20 23:08:19.000000000 +0200
> +++ linux/drivers/net/via-rhine.c 2004-05-20 23:11:26.000000000 +0200
> @@ -863,12 +863,12 @@
> if (np->default_port & 0x330) {
> /* FIXME: shouldn't someone check this variable? */
> /* np->medialock = 1; */
> - printk(KERN_INFO " Forcing %dMbs %s-duplex operation.\n",
> + printk(KERN_INFO " Forcing %dMbps %s-duplex operation.\n",
> (option & 0x300 ? 100 : 10),
> (option & 0x220 ? "full" : "half"));
> if (np->mii_cnt)
> mdio_write(dev, np->phys[0], MII_BMCR,
> - ((option & 0x300) ? 0x2000 : 0) | /* 100mbps? */
> + ((option & 0x300) ? 0x2000 : 0) | /* 100Mbps? */
> ((option & 0x220) ? 0x0100 : 0)); /* Full duplex? */
> }
> }