Subject: Re: Error messages in dmesg (regarding new IDE controller, new IDE drive)
To: Pete Rushmere <pete@rushmere.org>
From: Andy Ruhl <acruhl@gmail.com>
List: port-cobalt
Date: 09/16/2004 11:43:17
On Thu, 16 Sep 2004 19:11:00 +0100, Pete Rushmere <pete@rushmere.org> wrote:
> Carsten,
> 
>  >wd1: soft error (corrected)
> 
> Not sure about this one.... Maybe Andy has some ideas? Perhaps this is just
> due to the driver adjusting the DMA.. As said, the 40 pin cable seems to
> cure all the messages, but probably isn't the right way to do it - works
> for me though.. A hardware frig is easier than mucking about with the
> drivers I think...
> 
> My Qube has been running with no problems for about 6 months now, the disks
> take a severe hammering as it is my main server these days. I did need a
> quick re-install a few weeks ago, but I think I broke something ;-) Maybe I
> should leave it alone!
> 
> I'll be interested to hear if anyone else comes up with some ideas about
> this..... to be honest, I'm not much of a guru and kernel/driver hacking is
> a bit beyond me.
> 
> Kind Regards,
> Pete.

Wondering if I'm the Andy you're referring to...

This is a fairly common thing when IDE is downgrading itself to some
acceptable level. The dmesg will post what the drive is capable of,
but if the rest isn't then you'll see some messages. I doubt any of
this is a big deal unless you check your syslog and see some more of
these now and then during normal operation.

A 40 pin cable is a fine way to "fix" the problem. The driver will
pick it up, and the highest you can go is ATA33, which is all you're
getting anyway.

I've also seen problems like this when cabling isn't the best and you
get interference from other devices.

Andy