[ros-kernel] Atapi device detection bug

[Eric Kohl]

"Filip Navara" <xnavara at volny.cz> wrote:

> The previous code is illegal according to the ATA/ATAPI-6 standard. The
> device reset command is mandatory for ATAPI devices and calling it on
> ATA devices is strictly prohibited! I would accept any solution that
> will not send the DEVICE RESET command to ATA drives, be it software
> reset (as used by many other OSes) or something else.

Yes, that's true. The specification prohibits the use of the device reset
command but MS atapi.sys issues this command to ATA devices anyway. And one
of my books about IDE and SCSI interfaces says that the device reset command
is optional for ATA4-ATA6 devices.

I think we should stick with the soft reset method until we find a device
that doesn't work.

Btw, I'll apply my patch to atapi.c today.


Regards,
Eric