2006-09-21 00:42:06

by Ben Duncan

[permalink] [raw]
Subject: Request kernel 2.6.18 ..

Report: 2.6.18 has solved a lot of issues with
my 965 chipset Duo Core intel MB and has been
extremely stable ...

Now, any idea on a timeline for driver for the Marvell IDE
controller ?

Thanks ...

--
Ben Duncan - Business Network Solutions, Inc. 336 Elton Road Jackson MS, 39212
"Never attribute to malice, that which can be adequately explained by stupidity"
- Hanlon's Razor


2006-09-21 00:48:04

by Jeff Garzik

[permalink] [raw]
Subject: Re: Request kernel 2.6.18 ..

Ben Duncan wrote:
> Report: 2.6.18 has solved a lot of issues with
> my 965 chipset Duo Core intel MB and has been
> extremely stable ...
>
> Now, any idea on a timeline for driver for the Marvell IDE
> controller ?

No.

Ask Marvell... I have, to no avail.

Jeff



2006-09-21 09:10:01

by Alan

[permalink] [raw]
Subject: Re: Request kernel 2.6.18 ..

Ar Mer, 2006-09-20 am 19:31 -0500, ysgrifennodd Ben Duncan:
> Report: 2.6.18 has solved a lot of issues with
> my 965 chipset Duo Core intel MB and has been
> extremely stable ...
>
> Now, any idea on a timeline for driver for the Marvell IDE
> controller ?

Possibly never. Marvell don't currently seem to want to play. Now that
might be for several reasons, one of which is that its someone elses
chip rebadged. In that case someone has a chance of working out what it
copies (eg which bits change when you boot with or without a master or
slave on each channel is a good clue)

The current prognosis however is that you and/or other marvell users are
going to have to reverse-engineer the thing or just avoid boards using
that chip.

2006-09-21 09:37:52

by Avi Kivity

[permalink] [raw]
Subject: Re: Request kernel 2.6.18 ..

Alan Cox wrote:
>
> Ar Mer, 2006-09-20 am 19:31 -0500, ysgrifennodd Ben Duncan:
> > Report: 2.6.18 has solved a lot of issues with
> > my 965 chipset Duo Core intel MB and has been
> > extremely stable ...
> >
> > Now, any idea on a timeline for driver for the Marvell IDE
> > controller ?
>
> Possibly never. Marvell don't currently seem to want to play. Now that
> might be for several reasons, one of which is that its someone elses
> chip rebadged. In that case someone has a chance of working out what it
> copies (eg which bits change when you boot with or without a master or
> slave on each channel is a good clue)
>
> The current prognosis however is that you and/or other marvell users are
> going to have to reverse-engineer the thing or just avoid boards using
> that chip.
>

I have such a board, and all-generic-ide appears to work (at least
enough to load the installer from CD, all the rest is SATA). Perhaps it
can be quirked so the parameter won't be necessary?

--
error compiling committee.c: too many arguments to function

2006-09-21 21:36:20

by Jeff Garzik

[permalink] [raw]
Subject: Marvell PATA (was Re: Request kernel 2.6.18 ..)

Update: I have now have docs. Missed a few bits of info, but most of
it is there. Kudos to Marvell...

Jeff