Q: MTD RAM in OF Device Tree

Hi all,

is there a standard way to define a ram chip (in my case a battery-buffered=
 nv ram, attached to the 5200's Local Bus) in the OF tree, and are there an=
y drivers (on 2.6.29.1) which pick up the chip/partition specification for =
mtd?  Does such a driver exist, or do I have to write one (probably based o=
n plat-ram.c?)?  I think of something like

localbus {
=09[...]
=09ranges =3D < 0 0 0xfe000000 0x02000000
=09=09   1 0 0xfc000000 0x00080000>;

        flash@0,0 {
=09=09[flash definition & partitions]
=09};

=09nvram@1,0 {
=09=09compatible =3D "mtd-ram";
=09=09reg =3D <1 0x0 0x00080000>;
=09=09bank-width =3D <2>;
=09=09device-width =3D <2>;
=09=09#size-cells =3D <1>;
=09=09#address-cells =3D <1>;
=09=09part@0 {
=09=09=09label =3D "pdata";
=09=09=09reg =3D <0x00000000 0x00080000>;
=09=09};
=09};
};

Thanks,
Albrecht.

Arcor.de Gaming Area - kostenfrei daddeln bis der Arzt kommt!
Jetzt checken und aus =FCber 80 Spielen w=E4hlen!
http://www.arcor.de/footer-gaming/

Hello Albrecht,

(adding linux-mtd)

On Tue, May 19, 2009 at 01:59:00PM +0200, Albrecht Dreß wrote:

is there a standard way to define a ram chip (in my case a battery-buffered
nv ram, attached to the 5200's Local Bus) in the OF tree, and are there any
drivers (on 2.6.29.1) which pick up the chip/partition specification for mtd?
Does such a driver exist, or do I have to write one (probably based on
plat-ram.c?)?  I think of something like

I wrote such a driver (yet without partitioning support) and I am trying to get
it mainline, just didn't get any comments so far:

http://patchwork.ozlabs.org/patch/23557/
http://patchwork.ozlabs.org/patch/23556/

I'd be happy if you could give it a try and donate some
{Acked|Tested|Reviewed}-by tags. Maybe this will help for my next try to get
it mainline.

Regards,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

On Tue, May 19, 2009 at 6:27 AM, Wolfram Sang [off-list ref] wrote=
:

Hello Albrecht,

(adding linux-mtd)

On Tue, May 19, 2009 at 01:59:00PM +0200, Albrecht Dre=DF wrote:

quoted

is there a standard way to define a ram chip (in my case a battery-buffe=

red

quoted

nv ram, attached to the 5200's Local Bus) in the OF tree, and are there =

any

quoted

drivers (on 2.6.29.1) which pick up the chip/partition specification for=

 mtd?

quoted

Does such a driver exist, or do I have to write one (probably based on
plat-ram.c?)? =A0I think of something like

I wrote such a driver (yet without partitioning support) and I am trying =

to get

it mainline, just didn't get any comments so far:

http://patchwork.ozlabs.org/patch/23557/
http://patchwork.ozlabs.org/patch/23556/

I'd be happy if you could give it a try and donate some
{Acked|Tested|Reviewed}-by tags. Maybe this will help for my next try to =

get

it mainline.

I missed them when you posted them.  I'll go look now.
--=20
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.

Hi Wolfram!

Am 19.05.09 14:27 schrieb(en) Wolfram Sang:

I wrote such a driver (yet without partitioning support) and I am  
trying to get it mainline, just didn't get any comments so far:

http://patchwork.ozlabs.org/patch/23557/
http://patchwork.ozlabs.org/patch/23556/

Thanks a lot, this is *exactly* what I was looking for!  Actually,  
partition support is not important as I want to use the whole chip -  
512k isn't *that* much...

I'd be happy if you could give it a try and donate some  
{Acked|Tested|Reviewed}-by tags. Maybe this will help for my next try  
to get it mainline.

Will do that once I get it running completely.  The first test  
triggered some strange problems when I wrote to /dev/mtdn and read back  
the data (which is supposed to give "raw" direct access, right?) - the  
last dword in each 512 byte block is partly broken.  Probably something  
in the localbus layout or timing...

Thanks again,
Albrecht.