linuxdc-dev

[linuxdc-dev] Maple Bus magic number etc

[Adrian M. <ad...@mc...>]

Looking through the code for the driver I see this:

        /* Initialize hardware */
        ctrl_outl(MAPLE_MAGIC, MAPLE_RESET);
        ctrl_outl(0, MAPLE_RESET2);
        ctrl_outl(MAPLE_2MBPS|MAPLE_TIMEOUT(50000), MAPLE_SPEED);
=09ctrl_outl(PHYSADDR(maple_sendbuf), MAPLE_DMAADDR);
        ctrl_outl(1, MAPLE_ENABLE);

Is this documented anywhere? I cannot find anything on Marcus's site or o=
n=20
google.

Thanks

Adrian

[linuxdc-dev] Re: Maple Bus magic number etc

[Paul M. <le...@li...>]

On Mon, Jun 07, 2004 at 11:59:58PM +0100, Adrian McMenamin wrote:
> Looking through the code for the driver I see this:
>=20
>         /* Initialize hardware */
>         ctrl_outl(MAPLE_MAGIC, MAPLE_RESET);
>         ctrl_outl(0, MAPLE_RESET2);
>         ctrl_outl(MAPLE_2MBPS|MAPLE_TIMEOUT(50000), MAPLE_SPEED);
> 	ctrl_outl(PHYSADDR(maple_sendbuf), MAPLE_DMAADDR);
>         ctrl_outl(1, MAPLE_ENABLE);
>=20
> Is this documented anywhere? I cannot find anything on Marcus's site or o=
n=20
> google.
>=20
No, it's not documented, that's precisely why it's a "magic" value.

MAPLE_MAGIC is simply the "magic" value that needs to be written to MAPLE_R=
ESET
in order to reset the bus. I'm not aware of anyone having reverse engineered
the maple stuff any further then what documentation is readily available, s=
o it
will remain a magic value until someone figures out the register descriptio=
n.

Re: [linuxdc-dev] Re: Maple Bus magic number etc

[Adrian M. <ad...@mc...>]

On Tuesday 08 June 2004 00:19, Paul Mundt wrote:
> On Mon, Jun 07, 2004 at 11:59:58PM +0100, Adrian McMenamin wrote:
> > Looking through the code for the driver I see this:
> >
> >         /* Initialize hardware */
> >         ctrl_outl(MAPLE_MAGIC, MAPLE_RESET);
> >         ctrl_outl(0, MAPLE_RESET2);
> >         ctrl_outl(MAPLE_2MBPS|MAPLE_TIMEOUT(50000), MAPLE_SPEED);
> > =09ctrl_outl(PHYSADDR(maple_sendbuf), MAPLE_DMAADDR);
> >         ctrl_outl(1, MAPLE_ENABLE);
> >
> > Is this documented anywhere? I cannot find anything on Marcus's site =
or
> > on google.
>
> No, it's not documented, that's precisely why it's a "magic" value.
>
> MAPLE_MAGIC is simply the "magic" value that needs to be written to
> MAPLE_RESET in order to reset the bus. I'm not aware of anyone having
> reverse engineered the maple stuff any further then what documentation =
is
> readily available, so it will remain a magic value until someone figure=
s
> out the register description.

Would you know what signals these generate on the bus? Which ones produce=
=20
interrupts etc? Wondering which one to trap to initialise the devices to =
meet=20
the new driver specs.

About to do a bit of hacking to find I suppose.

Adrian