On Wednesday 15 October 2003 12:04 am, Gary Duzan wrote:
> In Message <20031014204453.GB1362@antioche.eu.org> ,
>    Manuel Bouyer <bouyer@antioche.eu.org> wrote:
>
> =>Sorry, I forgot another test. Please try attached patch.
>
>    That fixed it; I get wd0, wd1, and cd0, no problems. Thanks.

You probably don't realize this because you got lucky before, but the "ATAPI 
master with ATA slave" configuration that you're using is actually 
specifically unsupported in the ATA spec.  In this case the master is 
behaving incorrectly and causing you to get bogus results from the slave.

The proposed patch is really not correct, and it is not worthwhile to make 
many machines boot slower in order to support incorrect setups.