[ros-kernel] SCSI success

[Eric Kohl]

"Hartmut Birr" <Hartmut.Birr at gmx.de> wrote:

> Hi,
>
> with the last changes to the disk driver stack, I'm able to get working
six
> different scsi controller with there WinNT 4.0 drivers:
>
> - Adaptec - AHA2940UW
> - Initio - A100U2W (based on initio 1060)
> - Domex ? - DMX 3194 (based on initio 950)
> - Tekram - DC-390F (based on symbios logic 875)
> - Diamond Multimedia - fireport 40 (based on symbios logic 875)
> - ??? - MAX875UW (based on symbios logic 875)

Congratulations! This is a huge step ahead for ReactOS! :-)


> There was connected a disk, a cdrom drive and a cd burner to one or two of
> this controllers. Working means, I've compiled ros on ros. Ros and the
> compiler was located on the IDE disk and the ros source on the scsi disk.
> I've also tested if ros can boot from the scsi disk. It is possible, but
the
> bios order of the disk's (ide, scsi) and the load order of the drivers
must
> be the same.

We need to implement the service tags in order to get the load order right.


> It is also possible to install ros directly on the scsi disk,
> if the setupldr does load the correct drivers in the correct order. I've
> used a modified setupldr which loads the  drivers like atapi. It would be
> nice, if the setup can use a driver disk for such a controller.

I planned to implement oemdisk support but it was a waste of time until
recently because of the broken floppy driver. I'll start working on this as
soon as possible.


> While
> testing sometimes ros has scrambled a partition on the scsi disk (not with
> the current cvs), but only the partition which was directly accessed by
ros
> (booting or compiling). If anyone has a scsi system, he can test if ros
does
> work on it.

I'll try to test ReactOS with SCSI devices as soon as possible.


Regards,
Eric